all new onnx model for opencv

.gitignore

@@ -1 +1,3 @@
 ssd_mobilenet_v2_coco_2018_03_29/demo
+
+**/demo

efficientdet-d0/README.md

@@ -33,18 +33,22 @@ net = cv2.dnn.readNet("efficientdet-d0_2026jul.onnx")
 ```
 
 ### C++
-The C++ demo runs inference with OpenCV's DNN module. Adjust the OpenCV paths to your setup:
+The C++ demo runs inference with ONNX Runtime (C++ API) and uses OpenCV only for image I/O.
+Install ONNX Runtime (C++) from https://github.com/microsoft/onnxruntime/releases — this build
+uses `onnxruntime-linux-x64-1.25.0` — and adjust the ONNX Runtime and OpenCV paths to your setup:
 ```bash
+ORT=/path/to/onnxruntime-linux-x64-1.25.0    # ONNX Runtime release dir (contains include/ and lib/)
 OCV=/path/to/opencv                          # OpenCV source tree
 OCVBUILD=/path/to/opencv/build               # OpenCV build directory (generated headers + libs)
 g++ -std=c++17 demo.cpp -o demo \
+  -I$ORT/include \
   -I$OCV/include \
   -I$OCV/modules/core/include \
-  -I$OCV/modules/dnn/include \
   -I$OCV/modules/imgproc/include \
   -I$OCV/modules/imgcodecs/include \
   -I$OCVBUILD \
-  -L$OCVBUILD/lib -Wl,-rpath,$OCVBUILD/lib -lopencv_dnn -lopencv_imgcodecs -lopencv_imgproc -lopencv_core
+  -L$ORT/lib -Wl,-rpath,$ORT/lib -lonnxruntime \
+  -L$OCVBUILD/lib -Wl,-rpath,$OCVBUILD/lib -lopencv_imgcodecs -lopencv_imgproc -lopencv_core
 ./demo --model efficientdet-d0_2026jul.onnx --image example_outputs/input_image.png --output example_outputs/output_image.png
 ```
 

efficientdet-d0/demo.cpp

@@ -1,4 +1,4 @@
-#include <opencv2/dnn.hpp>
+#include <onnxruntime_cxx_api.h>
 #include <opencv2/imgproc.hpp>
 #include <opencv2/imgcodecs.hpp>
 #include <algorithm>
@@ -36,23 +36,40 @@ int main(int argc, char** argv)
     resize(rgb, rgb, Size(sz, sz));
     if (!rgb.isContinuous()) rgb = rgb.clone();
 
-    int blobShape[] = {1, sz, sz, 3};
-    Mat blob(4, blobShape, CV_8U, rgb.data);
-    dnn::Net net = dnn::readNetFromONNX(model);
-    net.setInput(blob);
-    std::vector<Mat> outs;
-    net.forward(outs, net.getUnconnectedOutLayersNames());
+    Ort::Env env(ORT_LOGGING_LEVEL_WARNING, "demo");
+    Ort::SessionOptions so;
+    Ort::Session session(env, model.c_str(), so);
+    Ort::AllocatorWithDefaultOptions alloc;
+
+    auto in_name = session.GetInputNameAllocated(0, alloc);
+    const char* in_names[] = {in_name.get()};
+
+    size_t nout = session.GetOutputCount();
+    std::vector<Ort::AllocatedStringPtr> out_holders;
+    std::vector<std::string> out_str;
+    for (size_t i = 0; i < nout; ++i)
+    {
+        out_holders.push_back(session.GetOutputNameAllocated(i, alloc));
+        out_str.push_back(out_holders.back().get());
+    }
+    std::vector<const char*> out_names;
+    for (auto& s : out_str) out_names.push_back(s.c_str());
+
+    std::array<int64_t, 4> shape = {1, sz, sz, 3};
+    auto mem = Ort::MemoryInfo::CreateCpu(OrtArenaAllocator, OrtMemTypeDefault);
+    Ort::Value input = Ort::Value::CreateTensor<uint8_t>(mem, rgb.data, (size_t)sz * sz * 3, shape.data(), shape.size());
+
+    auto outs = session.Run(Ort::RunOptions{nullptr}, in_names, &input, 1, out_names.data(), out_names.size());
 
     const float* boxp = nullptr;
     const float* clsp = nullptr;
     int n = 0, nc = 0;
     for (size_t i = 0; i < outs.size(); ++i)
     {
-        const Mat& o = outs[i];
-        const float* p = (const float*)o.data;
-        int last = o.size[o.dims - 1];
-        if (last == 4) { boxp = p; n = o.size[o.dims - 2]; }
-        else { clsp = p; nc = last; }
+        auto os = outs[i].GetTensorTypeAndShapeInfo().GetShape();
+        const float* p = outs[i].GetTensorMutableData<float>();
+        if (os.back() == 4) { boxp = p; n = (int)os[os.size() - 2]; }
+        else { clsp = p; nc = (int)os.back(); }
     }
 
     std::vector<std::array<float, 2>> baseWH;

efficientdet-d0/demo.py

@@ -4,6 +4,7 @@ import os
 
 import cv2 as cv
 import numpy as np
+import onnxruntime as ort
 
 here = os.path.dirname(os.path.abspath(__file__))
 sz = 512
@@ -52,10 +53,9 @@ def main():
     anchors = build_anchors()
     acx, acy, aw, ah = anchors[:, 0], anchors[:, 1], anchors[:, 2], anchors[:, 3]
 
-    net = cv.dnn.readNetFromONNX(model)
+    sess = ort.InferenceSession(model, providers=["CPUExecutionProvider"])
     inp = cv.resize(cv.cvtColor(img, cv.COLOR_BGR2RGB), (sz, sz))
-    net.setInput(inp[None].astype(np.uint8))
-    res = net.forward(net.getUnconnectedOutLayersNames())
+    res = sess.run(None, {sess.get_inputs()[0].name: inp[None].astype(np.uint8)})
     box = next(a for a in res if a.shape[-1] == 4).reshape(-1, 4)
     cls = next(a for a in res if a.shape[-1] != 4).reshape(box.shape[0], -1)
 

faster_rcnn_inception_v2_coco_2018_01_28/README.md

@@ -3,7 +3,7 @@
 Object detection with the Faster-RCNN meta-architecture and an Inception v2 backbone,
 trained on the COCO dataset. The model was originally distributed as a frozen TensorFlow
 graph (`faster_rcnn_inception_v2_coco_2018_01_28.pb`) from the TensorFlow Object Detection
-API and converted to ONNX for inference with OpenCV's DNN module.
+API and converted to ONNX for inference with ONNX Runtime.
 
 ## Model Details
 - **Architecture**: Faster-RCNN with an Inception v2 backbone
@@ -20,18 +20,22 @@ python demo.py --model faster_rcnn_inception_v2_coco_2018_01_28_2026jul.onnx --i
 ```
 
 ### C++
-The C++ demo runs inference with OpenCV's DNN module. Adjust the OpenCV paths to your setup:
+The C++ demo runs inference with ONNX Runtime (C++ API) and uses OpenCV only for image I/O.
+Install ONNX Runtime (C++) from https://github.com/microsoft/onnxruntime/releases — this build
+uses `onnxruntime-linux-x64-1.25.0` — and adjust the ONNX Runtime and OpenCV paths to your setup:
 ```bash
+ORT=/path/to/onnxruntime-linux-x64-1.25.0    # ONNX Runtime release dir (contains include/ and lib/)
 OCV=/path/to/opencv                          # OpenCV source tree
 OCVBUILD=/path/to/opencv/build               # OpenCV build directory (generated headers + libs)
 g++ -std=c++17 demo.cpp -o demo \
+  -I$ORT/include \
   -I$OCV/include \
   -I$OCV/modules/core/include \
-  -I$OCV/modules/dnn/include \
   -I$OCV/modules/imgproc/include \
   -I$OCV/modules/imgcodecs/include \
   -I$OCVBUILD \
-  -L$OCVBUILD/lib -Wl,-rpath,$OCVBUILD/lib -lopencv_dnn -lopencv_imgcodecs -lopencv_imgproc -lopencv_core
+  -L$ORT/lib -Wl,-rpath,$ORT/lib -lonnxruntime \
+  -L$OCVBUILD/lib -Wl,-rpath,$OCVBUILD/lib -lopencv_imgcodecs -lopencv_imgproc -lopencv_core
 ./demo --model faster_rcnn_inception_v2_coco_2018_01_28_2026jul.onnx --image example_outputs/input_image.png --output example_outputs/output_image.png
 ```
 

faster_rcnn_inception_v2_coco_2018_01_28/demo.cpp

@@ -1,4 +1,4 @@
-#include <opencv2/dnn.hpp>
+#include <onnxruntime_cxx_api.h>
 #include <opencv2/imgproc.hpp>
 #include <opencv2/imgcodecs.hpp>
 #include <array>
@@ -34,21 +34,38 @@ int main(int argc, char** argv)
     cv::resize(rgb, rgb, cv::Size(W, H));
     if (!rgb.isContinuous()) rgb = rgb.clone();
 
-    int blobShape[] = {1, H, W, 3};
-    cv::Mat blob(4, blobShape, CV_8U, rgb.data);
-    cv::dnn::Net net = cv::dnn::readNetFromONNX(model, cv::dnn::ENGINE_ORT);
-    net.setInput(blob);
-    std::vector<cv::String> out_strs = {"detection_boxes:0", "detection_scores:0", "detection_classes:0", "num_detections:0"};
-    std::vector<cv::Mat> outs;
-    net.forward(outs, out_strs);
+    Ort::Env env(ORT_LOGGING_LEVEL_WARNING, "demo");
+    Ort::SessionOptions so;
+    Ort::Session session(env, model.c_str(), so);
+    Ort::AllocatorWithDefaultOptions alloc;
+
+    auto in_name = session.GetInputNameAllocated(0, alloc);
+    const char* in_names[] = {in_name.get()};
+
+    size_t nout = session.GetOutputCount();
+    std::vector<Ort::AllocatedStringPtr> out_holders;
+    std::vector<std::string> out_strs;
+    for (size_t i = 0; i < nout; ++i)
+    {
+        out_holders.push_back(session.GetOutputNameAllocated(i, alloc));
+        out_strs.push_back(out_holders.back().get());
+    }
+    std::vector<const char*> out_names;
+    for (auto& s : out_strs) out_names.push_back(s.c_str());
+
+    std::array<int64_t, 4> shape = {1, H, W, 3};
+    auto mem = Ort::MemoryInfo::CreateCpu(OrtArenaAllocator, OrtMemTypeDefault);
+    Ort::Value input = Ort::Value::CreateTensor<uint8_t>(mem, rgb.data, (size_t)H * W * 3, shape.data(), shape.size());
+
+    auto outs = session.Run(Ort::RunOptions{nullptr}, in_names, &input, 1, out_names.data(), out_names.size());
 
     float* boxes = nullptr;
     float* scores = nullptr;
     float* classes = nullptr;
     float* numd = nullptr;
-    for (size_t i = 0; i < out_strs.size(); ++i)
+    for (size_t i = 0; i < nout; ++i)
     {
-        float* p = (float*)outs[i].data;
+        float* p = outs[i].GetTensorMutableData<float>();
         const std::string& n = out_strs[i];
         if (n.find("detection_boxes") != std::string::npos) boxes = p;
         else if (n.find("detection_scores") != std::string::npos) scores = p;

faster_rcnn_inception_v2_coco_2018_01_28/demo.py

@@ -4,6 +4,7 @@ import os
 
 import cv2 as cv
 import numpy as np
+import onnxruntime as ort
 
 here = os.path.dirname(os.path.abspath(__file__))
 
@@ -22,10 +23,9 @@ def main():
         raise SystemExit("could not read image: %s" % args.image)
 
     rgb = cv.resize(cv.cvtColor(img, cv.COLOR_BGR2RGB), (800, 600))
-    net = cv.dnn.readNetFromONNX(model, cv.dnn.ENGINE_ORT)
-    onames = ["detection_boxes:0", "detection_scores:0", "detection_classes:0", "num_detections:0"]
-    net.setInput(rgb[None].astype(np.uint8))
-    res = net.forward(onames)
+    sess = ort.InferenceSession(model, providers=["CPUExecutionProvider"])
+    res = sess.run(None, {sess.get_inputs()[0].name: rgb[None].astype(np.uint8)})
+    onames = [o.name for o in sess.get_outputs()]
     boxes = res[[i for i, n in enumerate(onames) if "detection_boxes" in n][0]].reshape(-1, 4)
     scores = res[[i for i, n in enumerate(onames) if "detection_scores" in n][0]].reshape(-1)
     classes = res[[i for i, n in enumerate(onames) if "detection_classes" in n][0]].reshape(-1)

faster_rcnn_resnet50_coco_2018_01_28/README.md

@@ -3,7 +3,7 @@
 Object detection with the Faster-RCNN meta-architecture and a ResNet-50 backbone,
 trained on the COCO dataset. The model was originally distributed as a frozen TensorFlow
 graph (`faster_rcnn_resnet50_coco_2018_01_28.pb`) from the TensorFlow Object Detection
-API and converted to ONNX for inference with OpenCV's DNN module.
+API and converted to ONNX for inference with ONNX Runtime.
 
 ## Model Details
 - **Architecture**: Faster-RCNN with a ResNet-50 backbone
@@ -20,18 +20,22 @@ python demo.py --model faster_rcnn_resnet50_coco_2018_01_28_2026jul.onnx --image
 ```
 
 ### C++
-The C++ demo runs inference with OpenCV's DNN module. Adjust the OpenCV paths to your setup:
+The C++ demo runs inference with ONNX Runtime (C++ API) and uses OpenCV only for image I/O.
+Install ONNX Runtime (C++) from https://github.com/microsoft/onnxruntime/releases — this build
+uses `onnxruntime-linux-x64-1.25.0` — and adjust the ONNX Runtime and OpenCV paths to your setup:
 ```bash
+ORT=/path/to/onnxruntime-linux-x64-1.25.0    # ONNX Runtime release dir (contains include/ and lib/)
 OCV=/path/to/opencv                          # OpenCV source tree
 OCVBUILD=/path/to/opencv/build               # OpenCV build directory (generated headers + libs)
 g++ -std=c++17 demo.cpp -o demo \
+  -I$ORT/include \
   -I$OCV/include \
   -I$OCV/modules/core/include \
-  -I$OCV/modules/dnn/include \
   -I$OCV/modules/imgproc/include \
   -I$OCV/modules/imgcodecs/include \
   -I$OCVBUILD \
-  -L$OCVBUILD/lib -Wl,-rpath,$OCVBUILD/lib -lopencv_dnn -lopencv_imgcodecs -lopencv_imgproc -lopencv_core
+  -L$ORT/lib -Wl,-rpath,$ORT/lib -lonnxruntime \
+  -L$OCVBUILD/lib -Wl,-rpath,$OCVBUILD/lib -lopencv_imgcodecs -lopencv_imgproc -lopencv_core
 ./demo --model faster_rcnn_resnet50_coco_2018_01_28_2026jul.onnx --image example_outputs/input_image.png --output example_outputs/output_image.png
 ```
 

faster_rcnn_resnet50_coco_2018_01_28/demo.cpp

@@ -1,4 +1,4 @@
-#include <opencv2/dnn.hpp>
+#include <onnxruntime_cxx_api.h>
 #include <opencv2/imgproc.hpp>
 #include <opencv2/imgcodecs.hpp>
 #include <array>
@@ -34,21 +34,38 @@ int main(int argc, char** argv)
     cv::resize(rgb, rgb, cv::Size(W, H));
     if (!rgb.isContinuous()) rgb = rgb.clone();
 
-    int blobShape[] = {1, H, W, 3};
-    cv::Mat blob(4, blobShape, CV_8U, rgb.data);
-    cv::dnn::Net net = cv::dnn::readNetFromONNX(model, cv::dnn::ENGINE_ORT);
-    net.setInput(blob);
-    std::vector<cv::String> out_strs = {"detection_boxes:0", "detection_scores:0", "detection_classes:0", "num_detections:0"};
-    std::vector<cv::Mat> outs;
-    net.forward(outs, out_strs);
+    Ort::Env env(ORT_LOGGING_LEVEL_WARNING, "demo");
+    Ort::SessionOptions so;
+    Ort::Session session(env, model.c_str(), so);
+    Ort::AllocatorWithDefaultOptions alloc;
+
+    auto in_name = session.GetInputNameAllocated(0, alloc);
+    const char* in_names[] = {in_name.get()};
+
+    size_t nout = session.GetOutputCount();
+    std::vector<Ort::AllocatedStringPtr> out_holders;
+    std::vector<std::string> out_strs;
+    for (size_t i = 0; i < nout; ++i)
+    {
+        out_holders.push_back(session.GetOutputNameAllocated(i, alloc));
+        out_strs.push_back(out_holders.back().get());
+    }
+    std::vector<const char*> out_names;
+    for (auto& s : out_strs) out_names.push_back(s.c_str());
+
+    std::array<int64_t, 4> shape = {1, H, W, 3};
+    auto mem = Ort::MemoryInfo::CreateCpu(OrtArenaAllocator, OrtMemTypeDefault);
+    Ort::Value input = Ort::Value::CreateTensor<uint8_t>(mem, rgb.data, (size_t)H * W * 3, shape.data(), shape.size());
+
+    auto outs = session.Run(Ort::RunOptions{nullptr}, in_names, &input, 1, out_names.data(), out_names.size());
 
     float* boxes = nullptr;
     float* scores = nullptr;
     float* classes = nullptr;
     float* numd = nullptr;
-    for (size_t i = 0; i < out_strs.size(); ++i)
+    for (size_t i = 0; i < nout; ++i)
     {
-        float* p = (float*)outs[i].data;
+        float* p = outs[i].GetTensorMutableData<float>();
         const std::string& n = out_strs[i];
         if (n.find("detection_boxes") != std::string::npos) boxes = p;
         else if (n.find("detection_scores") != std::string::npos) scores = p;

faster_rcnn_resnet50_coco_2018_01_28/demo.py

@@ -4,6 +4,7 @@ import os
 
 import cv2 as cv
 import numpy as np
+import onnxruntime as ort
 
 here = os.path.dirname(os.path.abspath(__file__))
 
@@ -22,10 +23,9 @@ def main():
         raise SystemExit("could not read image: %s" % args.image)
 
     rgb = cv.resize(cv.cvtColor(img, cv.COLOR_BGR2RGB), (800, 600))
-    net = cv.dnn.readNetFromONNX(model, cv.dnn.ENGINE_ORT)
-    onames = ["detection_boxes:0", "detection_scores:0", "detection_classes:0", "num_detections:0"]
-    net.setInput(rgb[None].astype(np.uint8))
-    res = net.forward(onames)
+    sess = ort.InferenceSession(model, providers=["CPUExecutionProvider"])
+    res = sess.run(None, {sess.get_inputs()[0].name: rgb[None].astype(np.uint8)})
+    onames = [o.name for o in sess.get_outputs()]
     boxes = res[[i for i, n in enumerate(onames) if "detection_boxes" in n][0]].reshape(-1, 4)
     scores = res[[i for i, n in enumerate(onames) if "detection_scores" in n][0]].reshape(-1)
     classes = res[[i for i, n in enumerate(onames) if "detection_classes" in n][0]].reshape(-1)

mask_rcnn_inception_v2_coco_2018_01_28/README.md

@@ -3,7 +3,7 @@
 Instance segmentation with the Mask-RCNN Inception v2 network trained on the COCO dataset.
 The model was originally distributed as a frozen TensorFlow graph
 (`mask_rcnn_inception_v2_coco_2018_01_28.pb`) from the TensorFlow Object Detection API
-and converted to ONNX for use with OpenCV's DNN module.
+and converted to ONNX for use with ONNX Runtime.
 
 ## Model Details
 - **Architecture**: Mask-RCNN with an Inception v2 backbone
@@ -22,18 +22,22 @@ python demo.py --model mask_rcnn_inception_v2_coco_2018_01_28_2026jul.onnx --ima
 ```
 
 ### C++
-The C++ demo runs inference with OpenCV's DNN module. Adjust the OpenCV paths to your setup:
+The C++ demo runs inference with ONNX Runtime (C++ API) and uses OpenCV only for image I/O.
+Install ONNX Runtime (C++) from https://github.com/microsoft/onnxruntime/releases — this build
+uses `onnxruntime-linux-x64-1.25.0` — and adjust the ONNX Runtime and OpenCV paths to your setup:
 ```bash
+ORT=/path/to/onnxruntime-linux-x64-1.25.0    # ONNX Runtime release dir (contains include/ and lib/)
 OCV=/path/to/opencv                          # OpenCV source tree
 OCVBUILD=/path/to/opencv/build               # OpenCV build directory (generated headers + libs)
 g++ -std=c++17 demo.cpp -o demo \
+  -I$ORT/include \
   -I$OCV/include \
   -I$OCV/modules/core/include \
-  -I$OCV/modules/dnn/include \
   -I$OCV/modules/imgproc/include \
   -I$OCV/modules/imgcodecs/include \
   -I$OCVBUILD \
-  -L$OCVBUILD/lib -Wl,-rpath,$OCVBUILD/lib -lopencv_dnn -lopencv_imgcodecs -lopencv_imgproc -lopencv_core
+  -L$ORT/lib -Wl,-rpath,$ORT/lib -lonnxruntime \
+  -L$OCVBUILD/lib -Wl,-rpath,$OCVBUILD/lib -lopencv_imgcodecs -lopencv_imgproc -lopencv_core
 ./demo --model mask_rcnn_inception_v2_coco_2018_01_28_2026jul.onnx --image example_outputs/input_image.png --output example_outputs/output_image.png
 ```
 

mask_rcnn_inception_v2_coco_2018_01_28/demo.cpp

@@ -1,4 +1,4 @@
-#include <opencv2/dnn.hpp>
+#include <onnxruntime_cxx_api.h>
 #include <opencv2/imgproc.hpp>
 #include <opencv2/imgcodecs.hpp>
 #include <algorithm>
@@ -35,22 +35,39 @@ int main(int argc, char** argv)
     cv::resize(rgb, rgb, cv::Size(W, H));
     if (!rgb.isContinuous()) rgb = rgb.clone();
 
-    int blobShape[] = {1, H, W, 3};
-    cv::Mat blob(4, blobShape, CV_8U, rgb.data);
-    cv::dnn::Net net = cv::dnn::readNetFromONNX(model, cv::dnn::ENGINE_ORT);
-    net.setInput(blob);
-    std::vector<cv::String> out_strs = {"num_detections:0", "detection_boxes:0", "detection_scores:0", "detection_classes:0", "detection_masks:0"};
-    std::vector<cv::Mat> outs;
-    net.forward(outs, out_strs);
+    Ort::Env env(ORT_LOGGING_LEVEL_WARNING, "demo");
+    Ort::SessionOptions so;
+    Ort::Session session(env, model.c_str(), so);
+    Ort::AllocatorWithDefaultOptions alloc;
+
+    auto in_name = session.GetInputNameAllocated(0, alloc);
+    const char* in_names[] = {in_name.get()};
+
+    size_t nout = session.GetOutputCount();
+    std::vector<Ort::AllocatedStringPtr> out_holders;
+    std::vector<std::string> out_strs;
+    for (size_t i = 0; i < nout; ++i)
+    {
+        out_holders.push_back(session.GetOutputNameAllocated(i, alloc));
+        out_strs.push_back(out_holders.back().get());
+    }
+    std::vector<const char*> out_names;
+    for (auto& s : out_strs) out_names.push_back(s.c_str());
+
+    std::array<int64_t, 4> shape = {1, H, W, 3};
+    auto mem = Ort::MemoryInfo::CreateCpu(OrtArenaAllocator, OrtMemTypeDefault);
+    Ort::Value input = Ort::Value::CreateTensor<uint8_t>(mem, rgb.data, (size_t)H * W * 3, shape.data(), shape.size());
+
+    auto outs = session.Run(Ort::RunOptions{nullptr}, in_names, &input, 1, out_names.data(), out_names.size());
 
     float* boxes = nullptr;
     float* scores = nullptr;
     float* classes = nullptr;
     float* numd = nullptr;
     float* masks = nullptr;
-    for (size_t i = 0; i < out_strs.size(); ++i)
+    for (size_t i = 0; i < nout; ++i)
     {
-        float* p = (float*)outs[i].data;
+        float* p = outs[i].GetTensorMutableData<float>();
         const std::string& n = out_strs[i];
         if (n.find("detection_boxes") != std::string::npos) boxes = p;
         else if (n.find("detection_scores") != std::string::npos) scores = p;

mask_rcnn_inception_v2_coco_2018_01_28/demo.py

@@ -4,12 +4,13 @@ import os
 
 import cv2 as cv
 import numpy as np
+import onnxruntime as ort
 
 here = os.path.dirname(os.path.abspath(__file__))
 
 
 def main():
-    parser = argparse.ArgumentParser(description="Mask-RCNN Inception v2 COCO (OpenCV DNN) detection + mask demo")
+    parser = argparse.ArgumentParser(description="Mask-RCNN Inception v2 COCO (ONNX Runtime) detection + mask demo")
     found = glob.glob(os.path.join(here, "*.onnx"))
     parser.add_argument("--model", default=found[0] if found else None)
     parser.add_argument("--image", default=os.path.join(here, "example_outputs", "input_image.png"))
@@ -21,11 +22,10 @@ def main():
     if img is None:
         raise SystemExit("could not read image: %s" % args.image)
 
-    net = cv.dnn.readNetFromONNX(args.model, cv.dnn.ENGINE_ORT)
+    sess = ort.InferenceSession(args.model, providers=["CPUExecutionProvider"])
     rgb = cv.resize(cv.cvtColor(img, cv.COLOR_BGR2RGB), (800, 800))
-    onames = ["num_detections:0", "detection_boxes:0", "detection_scores:0", "detection_classes:0", "detection_masks:0"]
-    net.setInput(rgb[None].astype(np.uint8))
-    res = net.forward(onames)
+    res = sess.run(None, {sess.get_inputs()[0].name: rgb[None].astype(np.uint8)})
+    onames = [o.name for o in sess.get_outputs()]
     boxes = res[[i for i, n in enumerate(onames) if "detection_boxes" in n][0]].reshape(-1, 4)
     scores = res[[i for i, n in enumerate(onames) if "detection_scores" in n][0]].reshape(-1)
     classes = res[[i for i, n in enumerate(onames) if "detection_classes" in n][0]].reshape(-1)

opencv_face_detector_uint8/README.md

@@ -2,7 +2,7 @@
 
 Single-shot face detection with the OpenCV SSD ResNet-10 network. The model ships in the
 OpenCV project as a quantized frozen TensorFlow graph (`opencv_face_detector_uint8.pb`) and is
-converted here to ONNX for use with OpenCV's DNN module. Only the backbone is
+converted here to ONNX for use with OpenCV's DNN module and onnxruntime. Only the backbone is
 exported — PriorBox generation, the confidence softmax, variance decode, score threshold and NMS
 are run in host code (see `demo.py` / `demo.cpp`).
 
@@ -36,18 +36,22 @@ net = cv2.dnn.readNet("opencv_face_detector_uint8_2026jul.onnx")
 ```
 
 ### C++
-The C++ demo runs inference with OpenCV's DNN module. Adjust the OpenCV paths to your setup:
+The C++ demo runs inference with ONNX Runtime (C++ API) and uses OpenCV only for image I/O.
+Install ONNX Runtime (C++) from https://github.com/microsoft/onnxruntime/releases — this build
+uses `onnxruntime-linux-x64-1.25.0` — and adjust the ONNX Runtime and OpenCV paths to your setup:
 ```bash
+ORT=/path/to/onnxruntime-linux-x64-1.25.0    # ONNX Runtime release dir (contains include/ and lib/)
 OCV=/path/to/opencv                          # OpenCV source tree
 OCVBUILD=/path/to/opencv/build               # OpenCV build directory (generated headers + libs)
 g++ -std=c++17 demo.cpp -o demo \
+  -I$ORT/include \
   -I$OCV/include \
   -I$OCV/modules/core/include \
-  -I$OCV/modules/dnn/include \
   -I$OCV/modules/imgproc/include \
   -I$OCV/modules/imgcodecs/include \
   -I$OCVBUILD \
-  -L$OCVBUILD/lib -Wl,-rpath,$OCVBUILD/lib -lopencv_dnn -lopencv_imgcodecs -lopencv_imgproc -lopencv_core
+  -L$ORT/lib -Wl,-rpath,$ORT/lib -lonnxruntime \
+  -L$OCVBUILD/lib -Wl,-rpath,$OCVBUILD/lib -lopencv_imgcodecs -lopencv_imgproc -lopencv_core
 ./demo --model opencv_face_detector_uint8_2026jul.onnx --image example_outputs/input_image.png --output example_outputs/output_image.png
 ```
 

opencv_face_detector_uint8/demo.cpp

@@ -1,4 +1,4 @@
-#include <opencv2/dnn.hpp>
+#include <onnxruntime_cxx_api.h>
 #include <opencv2/imgproc.hpp>
 #include <opencv2/imgcodecs.hpp>
 #include <algorithm>
@@ -40,20 +40,39 @@ int main(int argc, char** argv)
     subtract(inp, Scalar(104, 177, 123), inp);
     if (!inp.isContinuous()) inp = inp.clone();
 
-    int blobShape[] = {1, sz, sz, 3};
-    Mat blob(4, blobShape, CV_32F, inp.data);
-    dnn::Net net = dnn::readNetFromONNX(model);
-    net.setInput(blob);
-    std::vector<Mat> outs;
-    net.forward(outs, net.getUnconnectedOutLayersNames());
+    Ort::Env env(ORT_LOGGING_LEVEL_WARNING, "demo");
+    Ort::SessionOptions so;
+    Ort::Session session(env, model.c_str(), so);
+    Ort::AllocatorWithDefaultOptions alloc;
+
+    auto in_name = session.GetInputNameAllocated(0, alloc);
+    const char* in_names[] = {in_name.get()};
+
+    size_t nout = session.GetOutputCount();
+    std::vector<Ort::AllocatedStringPtr> out_holders;
+    std::vector<std::string> out_str;
+    for (size_t i = 0; i < nout; ++i)
+    {
+        out_holders.push_back(session.GetOutputNameAllocated(i, alloc));
+        out_str.push_back(out_holders.back().get());
+    }
+    std::vector<const char*> out_names;
+    for (auto& s : out_str) out_names.push_back(s.c_str());
+
+    std::array<int64_t, 4> shape = {1, sz, sz, 3};
+    auto mem = Ort::MemoryInfo::CreateCpu(OrtArenaAllocator, OrtMemTypeDefault);
+    Ort::Value input = Ort::Value::CreateTensor<float>(mem, (float*)inp.data, (size_t)sz * sz * 3, shape.data(), shape.size());
+
+    auto outs = session.Run(Ort::RunOptions{nullptr}, in_names, &input, 1, out_names.data(), out_names.size());
 
     const float* loc = nullptr;
     const float* conf = nullptr;
     for (size_t i = 0; i < outs.size(); ++i)
     {
-        const Mat& o = outs[i];
-        size_t tot = o.total();
-        const float* p = (const float*)o.data;
+        auto os = outs[i].GetTensorTypeAndShapeInfo().GetShape();
+        size_t tot = 1;
+        for (auto d : os) tot *= (size_t)d;
+        const float* p = outs[i].GetTensorMutableData<float>();
         if (tot == 35568) loc = p;
         else if (tot == 17784) conf = p;
     }

opencv_face_detector_uint8/demo.py

@@ -4,6 +4,7 @@ import os
 
 import cv2 as cv
 import numpy as np
+import onnxruntime as ort
 
 here = os.path.dirname(os.path.abspath(__file__))
 
@@ -56,10 +57,9 @@ def main():
 
     inp = cv.resize(img, (sz, sz)).astype(np.float32) - np.array([104.0, 177.0, 123.0], np.float32)
 
-    net = cv.dnn.readNetFromONNX(args.model)
-    onames = net.getUnconnectedOutLayersNames()
-    net.setInput(inp[None])
-    res = net.forward(onames)
+    sess = ort.InferenceSession(args.model, providers=["CPUExecutionProvider"])
+    res = sess.run(None, {sess.get_inputs()[0].name: inp[None]})
+    onames = [o.name for o in sess.get_outputs()]
     loc = res[[i for i, n in enumerate(onames) if "mbox_loc" in n][0]].reshape(-1, 4)
     conf = res[[i for i, n in enumerate(onames) if "mbox_conf" in n][0]].reshape(-1, 2)
 

ssd_inception_v2_coco_2017_11_17/README.md

@@ -19,18 +19,22 @@ python demo.py --model ssd_inception_v2_coco_2017_11_17_2026jul.onnx --image exa
 ```
 
 ### C++
-The C++ demo runs inference with OpenCV's DNN module. Adjust the OpenCV paths to your setup:
+The C++ demo runs inference with ONNX Runtime (C++ API) and uses OpenCV only for image I/O.
+Install ONNX Runtime (C++) from https://github.com/microsoft/onnxruntime/releases — this build
+uses `onnxruntime-linux-x64-1.25.0` — and adjust the ONNX Runtime and OpenCV paths to your setup:
 ```bash
+ORT=/path/to/onnxruntime-linux-x64-1.25.0    # ONNX Runtime release dir (contains include/ and lib/)
 OCV=/path/to/opencv                          # OpenCV source tree
 OCVBUILD=/path/to/opencv/build               # OpenCV build directory (generated headers + libs)
 g++ -std=c++17 demo.cpp -o demo \
+  -I$ORT/include \
   -I$OCV/include \
   -I$OCV/modules/core/include \
-  -I$OCV/modules/dnn/include \
   -I$OCV/modules/imgproc/include \
   -I$OCV/modules/imgcodecs/include \
   -I$OCVBUILD \
-  -L$OCVBUILD/lib -Wl,-rpath,$OCVBUILD/lib -lopencv_dnn -lopencv_imgcodecs -lopencv_imgproc -lopencv_core
+  -L$ORT/lib -Wl,-rpath,$ORT/lib -lonnxruntime \
+  -L$OCVBUILD/lib -Wl,-rpath,$OCVBUILD/lib -lopencv_imgcodecs -lopencv_imgproc -lopencv_core
 ./demo --model ssd_inception_v2_coco_2017_11_17_2026jul.onnx --image example_outputs/input_image.png --output example_outputs/output_image.png
 ```
 

ssd_inception_v2_coco_2017_11_17/demo.cpp

@@ -1,4 +1,4 @@
-#include <opencv2/dnn.hpp>
+#include <onnxruntime_cxx_api.h>
 #include <opencv2/imgproc.hpp>
 #include <opencv2/imgcodecs.hpp>
 #include <array>
@@ -35,22 +35,39 @@ int main(int argc, char** argv)
     resize(rgb, rgb, Size(300, 300));
     if (!rgb.isContinuous()) rgb = rgb.clone();
 
-    int blobShape[] = {1, 300, 300, 3};
-    Mat blob(4, blobShape, CV_8U, rgb.data);
-    dnn::Net net = dnn::readNetFromONNX(model, dnn::ENGINE_ORT);
-    net.setInput(blob);
-    std::vector<String> out_str = {"detection_boxes:0", "detection_scores:0", "detection_classes:0", "num_detections:0"};
-    std::vector<Mat> outs;
-    net.forward(outs, out_str);
+    Ort::Env env(ORT_LOGGING_LEVEL_WARNING, "demo");
+    Ort::SessionOptions so;
+    Ort::Session session(env, model.c_str(), so);
+    Ort::AllocatorWithDefaultOptions alloc;
+
+    auto in_name = session.GetInputNameAllocated(0, alloc);
+    const char* in_names[] = {in_name.get()};
+
+    size_t out_count = session.GetOutputCount();
+    std::vector<Ort::AllocatedStringPtr> out_holders;
+    std::vector<std::string> out_str;
+    std::vector<const char*> out_names;
+    for (size_t i = 0; i < out_count; ++i)
+    {
+        out_holders.push_back(session.GetOutputNameAllocated(i, alloc));
+        out_str.push_back(out_holders.back().get());
+        out_names.push_back(out_str.back().c_str());
+    }
+
+    std::array<int64_t, 4> shape = {1, 300, 300, 3};
+    auto mem = Ort::MemoryInfo::CreateCpu(OrtArenaAllocator, OrtMemTypeDefault);
+    Ort::Value input = Ort::Value::CreateTensor<uint8_t>(mem, rgb.data, 300 * 300 * 3, shape.data(), shape.size());
+
+    auto outs = session.Run(Ort::RunOptions{nullptr}, in_names, &input, 1, out_names.data(), out_names.size());
 
     const float *boxes = 0, *scores = 0, *classes = 0, *num = 0;
-    for (size_t i = 0; i < out_str.size(); ++i)
+    for (size_t i = 0; i < out_count; ++i)
     {
         const std::string& n = out_str[i];
-        if (n.find("detection_boxes") != std::string::npos) boxes = (const float*)outs[i].data;
-        else if (n.find("detection_scores") != std::string::npos) scores = (const float*)outs[i].data;
-        else if (n.find("detection_classes") != std::string::npos) classes = (const float*)outs[i].data;
-        else if (n.find("num_detections") != std::string::npos) num = (const float*)outs[i].data;
+        if (n.find("detection_boxes") != std::string::npos) boxes = outs[i].GetTensorMutableData<float>();
+        else if (n.find("detection_scores") != std::string::npos) scores = outs[i].GetTensorMutableData<float>();
+        else if (n.find("detection_classes") != std::string::npos) classes = outs[i].GetTensorMutableData<float>();
+        else if (n.find("num_detections") != std::string::npos) num = outs[i].GetTensorMutableData<float>();
     }
     if (!boxes || !scores || !classes || !num)
     {

ssd_inception_v2_coco_2017_11_17/demo.py

@@ -4,6 +4,7 @@ import os
 
 import cv2 as cv
 import numpy as np
+import onnxruntime as ort
 
 here = os.path.dirname(os.path.abspath(__file__))
 
@@ -22,10 +23,9 @@ def main():
 
     rgb = cv.resize(cv.cvtColor(img, cv.COLOR_BGR2RGB), (300, 300))
 
-    net = cv.dnn.readNetFromONNX(args.model, cv.dnn.ENGINE_ORT)
-    onames = ["detection_boxes:0", "detection_scores:0", "detection_classes:0", "num_detections:0"]
-    net.setInput(rgb[None].astype(np.uint8))
-    res = net.forward(onames)
+    sess = ort.InferenceSession(args.model, providers=["CPUExecutionProvider"])
+    res = sess.run(None, {sess.get_inputs()[0].name: rgb[None].astype(np.uint8)})
+    onames = [o.name for o in sess.get_outputs()]
     boxes = res[[i for i, n in enumerate(onames) if "detection_boxes" in n][0]].reshape(-1, 4)
     scores = res[[i for i, n in enumerate(onames) if "detection_scores" in n][0]].reshape(-1)
     classes = res[[i for i, n in enumerate(onames) if "detection_classes" in n][0]].reshape(-1)

ssd_mobilenet_v1_ppn_coco/README.md

@@ -20,18 +20,22 @@ python demo.py --model ssd_mobilenet_v1_ppn_coco_2026jul.onnx --image example_ou
 ```
 
 ### C++
-The C++ demo runs inference with OpenCV's DNN module. Adjust the OpenCV paths to your setup:
+The C++ demo runs inference with ONNX Runtime (C++ API) and uses OpenCV only for image I/O.
+Install ONNX Runtime (C++) from https://github.com/microsoft/onnxruntime/releases — this build
+uses `onnxruntime-linux-x64-1.25.0` — and adjust the ONNX Runtime and OpenCV paths to your setup:
 ```bash
+ORT=/path/to/onnxruntime-linux-x64-1.25.0    # ONNX Runtime release dir (contains include/ and lib/)
 OCV=/path/to/opencv                          # OpenCV source tree
 OCVBUILD=/path/to/opencv/build               # OpenCV build directory (generated headers + libs)
 g++ -std=c++17 demo.cpp -o demo \
+  -I$ORT/include \
   -I$OCV/include \
   -I$OCV/modules/core/include \
-  -I$OCV/modules/dnn/include \
   -I$OCV/modules/imgproc/include \
   -I$OCV/modules/imgcodecs/include \
   -I$OCVBUILD \
-  -L$OCVBUILD/lib -Wl,-rpath,$OCVBUILD/lib -lopencv_dnn -lopencv_imgcodecs -lopencv_imgproc -lopencv_core
+  -L$ORT/lib -Wl,-rpath,$ORT/lib -lonnxruntime \
+  -L$OCVBUILD/lib -Wl,-rpath,$OCVBUILD/lib -lopencv_imgcodecs -lopencv_imgproc -lopencv_core
 ./demo --model ssd_mobilenet_v1_ppn_coco_2026jul.onnx --image example_outputs/input_image.png --output example_outputs/output_image.png
 ```
 

ssd_mobilenet_v1_ppn_coco/demo.cpp

@@ -1,4 +1,4 @@
-#include <opencv2/dnn.hpp>
+#include <onnxruntime_cxx_api.h>
 #include <opencv2/imgproc.hpp>
 #include <opencv2/imgcodecs.hpp>
 #include <array>
@@ -35,22 +35,39 @@ int main(int argc, char** argv)
     resize(rgb, rgb, Size(300, 300));
     if (!rgb.isContinuous()) rgb = rgb.clone();
 
-    int blobShape[] = {1, 300, 300, 3};
-    Mat blob(4, blobShape, CV_8U, rgb.data);
-    dnn::Net net = dnn::readNetFromONNX(model, dnn::ENGINE_ORT);
-    net.setInput(blob);
-    std::vector<String> out_str = {"detection_boxes:0", "detection_scores:0", "detection_classes:0", "num_detections:0"};
-    std::vector<Mat> outs;
-    net.forward(outs, out_str);
+    Ort::Env env(ORT_LOGGING_LEVEL_WARNING, "demo");
+    Ort::SessionOptions so;
+    Ort::Session session(env, model.c_str(), so);
+    Ort::AllocatorWithDefaultOptions alloc;
+
+    auto in_name = session.GetInputNameAllocated(0, alloc);
+    const char* in_names[] = {in_name.get()};
+
+    size_t out_count = session.GetOutputCount();
+    std::vector<Ort::AllocatedStringPtr> out_holders;
+    std::vector<std::string> out_str;
+    std::vector<const char*> out_names;
+    for (size_t i = 0; i < out_count; ++i)
+    {
+        out_holders.push_back(session.GetOutputNameAllocated(i, alloc));
+        out_str.push_back(out_holders.back().get());
+        out_names.push_back(out_str.back().c_str());
+    }
+
+    std::array<int64_t, 4> shape = {1, 300, 300, 3};
+    auto mem = Ort::MemoryInfo::CreateCpu(OrtArenaAllocator, OrtMemTypeDefault);
+    Ort::Value input = Ort::Value::CreateTensor<uint8_t>(mem, rgb.data, 300 * 300 * 3, shape.data(), shape.size());
+
+    auto outs = session.Run(Ort::RunOptions{nullptr}, in_names, &input, 1, out_names.data(), out_names.size());
 
     const float *boxes = 0, *scores = 0, *classes = 0, *num = 0;
-    for (size_t i = 0; i < out_str.size(); ++i)
+    for (size_t i = 0; i < out_count; ++i)
     {
         const std::string& n = out_str[i];
-        if (n.find("detection_boxes") != std::string::npos) boxes = (const float*)outs[i].data;
-        else if (n.find("detection_scores") != std::string::npos) scores = (const float*)outs[i].data;
-        else if (n.find("detection_classes") != std::string::npos) classes = (const float*)outs[i].data;
-        else if (n.find("num_detections") != std::string::npos) num = (const float*)outs[i].data;
+        if (n.find("detection_boxes") != std::string::npos) boxes = outs[i].GetTensorMutableData<float>();
+        else if (n.find("detection_scores") != std::string::npos) scores = outs[i].GetTensorMutableData<float>();
+        else if (n.find("detection_classes") != std::string::npos) classes = outs[i].GetTensorMutableData<float>();
+        else if (n.find("num_detections") != std::string::npos) num = outs[i].GetTensorMutableData<float>();
     }
     if (!boxes || !scores || !classes || !num)
     {

ssd_mobilenet_v1_ppn_coco/demo.py

@@ -4,6 +4,7 @@ import os
 
 import cv2 as cv
 import numpy as np
+import onnxruntime as ort
 
 here = os.path.dirname(os.path.abspath(__file__))
 
@@ -22,10 +23,9 @@ def main():
 
     rgb = cv.resize(cv.cvtColor(img, cv.COLOR_BGR2RGB), (300, 300))
 
-    net = cv.dnn.readNetFromONNX(args.model, cv.dnn.ENGINE_ORT)
-    onames = ["detection_boxes:0", "detection_scores:0", "detection_classes:0", "num_detections:0"]
-    net.setInput(rgb[None].astype(np.uint8))
-    res = net.forward(onames)
+    sess = ort.InferenceSession(args.model, providers=["CPUExecutionProvider"])
+    res = sess.run(None, {sess.get_inputs()[0].name: rgb[None].astype(np.uint8)})
+    onames = [o.name for o in sess.get_outputs()]
     boxes = res[[i for i, n in enumerate(onames) if "detection_boxes" in n][0]].reshape(-1, 4)
     scores = res[[i for i, n in enumerate(onames) if "detection_scores" in n][0]].reshape(-1)
     classes = res[[i for i, n in enumerate(onames) if "detection_classes" in n][0]].reshape(-1)

ssd_mobilenet_v2_coco_2018_03_29/LICENSE

@@ -0,0 +1,212 @@
+Copyright 2022 Google LLC. All rights reserved.
+
+All files in the following folders:
+/community
+/official
+/orbit
+/research
+/tensorflow_models
+
+Are licensed as follows:
+
+                                 Apache License
+                           Version 2.0, January 2004
+                        http://www.apache.org/licenses/
+
+   TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION
+
+   1. Definitions.
+
+      "License" shall mean the terms and conditions for use, reproduction,
+      and distribution as defined by Sections 1 through 9 of this document.
+
+      "Licensor" shall mean the copyright owner or entity authorized by
+      the copyright owner that is granting the License.
+
+      "Legal Entity" shall mean the union of the acting entity and all
+      other entities that control, are controlled by, or are under common
+      control with that entity. For the purposes of this definition,
+      "control" means (i) the power, direct or indirect, to cause the
+      direction or management of such entity, whether by contract or
+      otherwise, or (ii) ownership of fifty percent (50%) or more of the
+      outstanding shares, or (iii) beneficial ownership of such entity.
+
+      "You" (or "Your") shall mean an individual or Legal Entity
+      exercising permissions granted by this License.
+
+      "Source" form shall mean the preferred form for making modifications,
+      including but not limited to software source code, documentation
+      source, and configuration files.
+
+      "Object" form shall mean any form resulting from mechanical
+      transformation or translation of a Source form, including but
+      not limited to compiled object code, generated documentation,
+      and conversions to other media types.
+
+      "Work" shall mean the work of authorship, whether in Source or
+      Object form, made available under the License, as indicated by a
+      copyright notice that is included in or attached to the work
+      (an example is provided in the Appendix below).
+
+      "Derivative Works" shall mean any work, whether in Source or Object
+      form, that is based on (or derived from) the Work and for which the
+      editorial revisions, annotations, elaborations, or other modifications
+      represent, as a whole, an original work of authorship. For the purposes
+      of this License, Derivative Works shall not include works that remain
+      separable from, or merely link (or bind by name) to the interfaces of,
+      the Work and Derivative Works thereof.
+
+      "Contribution" shall mean any work of authorship, including
+      the original version of the Work and any modifications or additions
+      to that Work or Derivative Works thereof, that is intentionally
+      submitted to Licensor for inclusion in the Work by the copyright owner
+      or by an individual or Legal Entity authorized to submit on behalf of
+      the copyright owner. For the purposes of this definition, "submitted"
+      means any form of electronic, verbal, or written communication sent
+      to the Licensor or its representatives, including but not limited to
+      communication on electronic mailing lists, source code control systems,
+      and issue tracking systems that are managed by, or on behalf of, the
+      Licensor for the purpose of discussing and improving the Work, but
+      excluding communication that is conspicuously marked or otherwise
+      designated in writing by the copyright owner as "Not a Contribution."
+
+      "Contributor" shall mean Licensor and any individual or Legal Entity
+      on behalf of whom a Contribution has been received by Licensor and
+      subsequently incorporated within the Work.
+
+   2. Grant of Copyright License. Subject to the terms and conditions of
+      this License, each Contributor hereby grants to You a perpetual,
+      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+      copyright license to reproduce, prepare Derivative Works of,
+      publicly display, publicly perform, sublicense, and distribute the
+      Work and such Derivative Works in Source or Object form.
+
+   3. Grant of Patent License. Subject to the terms and conditions of
+      this License, each Contributor hereby grants to You a perpetual,
+      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+      (except as stated in this section) patent license to make, have made,
+      use, offer to sell, sell, import, and otherwise transfer the Work,
+      where such license applies only to those patent claims licensable
+      by such Contributor that are necessarily infringed by their
+      Contribution(s) alone or by combination of their Contribution(s)
+      with the Work to which such Contribution(s) was submitted. If You
+      institute patent litigation against any entity (including a
+      cross-claim or counterclaim in a lawsuit) alleging that the Work
+      or a Contribution incorporated within the Work constitutes direct
+      or contributory patent infringement, then any patent licenses
+      granted to You under this License for that Work shall terminate
+      as of the date such litigation is filed.
+
+   4. Redistribution. You may reproduce and distribute copies of the
+      Work or Derivative Works thereof in any medium, with or without
+      modifications, and in Source or Object form, provided that You
+      meet the following conditions:
+
+      (a) You must give any other recipients of the Work or
+          Derivative Works a copy of this License; and
+
+      (b) You must cause any modified files to carry prominent notices
+          stating that You changed the files; and
+
+      (c) You must retain, in the Source form of any Derivative Works
+          that You distribute, all copyright, patent, trademark, and
+          attribution notices from the Source form of the Work,
+          excluding those notices that do not pertain to any part of
+          the Derivative Works; and
+
+      (d) If the Work includes a "NOTICE" text file as part of its
+          distribution, then any Derivative Works that You distribute must
+          include a readable copy of the attribution notices contained
+          within such NOTICE file, excluding those notices that do not
+          pertain to any part of the Derivative Works, in at least one
+          of the following places: within a NOTICE text file distributed
+          as part of the Derivative Works; within the Source form or
+          documentation, if provided along with the Derivative Works; or,
+          within a display generated by the Derivative Works, if and
+          wherever such third-party notices normally appear. The contents
+          of the NOTICE file are for informational purposes only and
+          do not modify the License. You may add Your own attribution
+          notices within Derivative Works that You distribute, alongside
+          or as an addendum to the NOTICE text from the Work, provided
+          that such additional attribution notices cannot be construed
+          as modifying the License.
+
+      You may add Your own copyright statement to Your modifications and
+      may provide additional or different license terms and conditions
+      for use, reproduction, or distribution of Your modifications, or
+      for any such Derivative Works as a whole, provided Your use,
+      reproduction, and distribution of the Work otherwise complies with
+      the conditions stated in this License.
+
+   5. Submission of Contributions. Unless You explicitly state otherwise,
+      any Contribution intentionally submitted for inclusion in the Work
+      by You to the Licensor shall be under the terms and conditions of
+      this License, without any additional terms or conditions.
+      Notwithstanding the above, nothing herein shall supersede or modify
+      the terms of any separate license agreement you may have executed
+      with Licensor regarding such Contributions.
+
+   6. Trademarks. This License does not grant permission to use the trade
+      names, trademarks, service marks, or product names of the Licensor,
+      except as required for reasonable and customary use in describing the
+      origin of the Work and reproducing the content of the NOTICE file.
+
+   7. Disclaimer of Warranty. Unless required by applicable law or
+      agreed to in writing, Licensor provides the Work (and each
+      Contributor provides its Contributions) on an "AS IS" BASIS,
+      WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
+      implied, including, without limitation, any warranties or conditions
+      of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A
+      PARTICULAR PURPOSE. You are solely responsible for determining the
+      appropriateness of using or redistributing the Work and assume any
+      risks associated with Your exercise of permissions under this License.
+
+   8. Limitation of Liability. In no event and under no legal theory,
+      whether in tort (including negligence), contract, or otherwise,
+      unless required by applicable law (such as deliberate and grossly
+      negligent acts) or agreed to in writing, shall any Contributor be
+      liable to You for damages, including any direct, indirect, special,
+      incidental, or consequential damages of any character arising as a
+      result of this License or out of the use or inability to use the
+      Work (including but not limited to damages for loss of goodwill,
+      work stoppage, computer failure or malfunction, or any and all
+      other commercial damages or losses), even if such Contributor
+      has been advised of the possibility of such damages.
+
+   9. Accepting Warranty or Additional Liability. While redistributing
+      the Work or Derivative Works thereof, You may choose to offer,
+      and charge a fee for, acceptance of support, warranty, indemnity,
+      or other liability obligations and/or rights consistent with this
+      License. However, in accepting such obligations, You may act only
+      on Your own behalf and on Your sole responsibility, not on behalf
+      of any other Contributor, and only if You agree to indemnify,
+      defend, and hold each Contributor harmless for any liability
+      incurred by, or claims asserted against, such Contributor by reason
+      of your accepting any such warranty or additional liability.
+
+   END OF TERMS AND CONDITIONS
+
+   APPENDIX: How to apply the Apache License to your work.
+
+      To apply the Apache License to your work, attach the following
+      boilerplate notice, with the fields enclosed by brackets "[]"
+      replaced with your own identifying information. (Don't include
+      the brackets!)  The text should be enclosed in the appropriate
+      comment syntax for the file format. We also recommend that a
+      file or class name and description of purpose be included on the
+      same "printed page" as the copyright notice for easier
+      identification within third-party archives.
+
+   Copyright 2016, The Authors.
+
+   Licensed under the Apache License, Version 2.0 (the "License");
+   you may not use this file except in compliance with the License.
+   You may obtain a copy of the License at
+
+       http://www.apache.org/licenses/LICENSE-2.0
+
+   Unless required by applicable law or agreed to in writing, software
+   distributed under the License is distributed on an "AS IS" BASIS,
+   WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+   See the License for the specific language governing permissions and
+   limitations under the License.

ssd_mobilenet_v2_coco_2018_03_29/README.md

@@ -0,0 +1,52 @@
+# SSD MobileNet v2 COCO
+
+Object detection with a Single Shot MultiBox Detector (SSD) built on a MobileNet v2 backbone,
+trained on the COCO dataset. The model was originally distributed as a frozen TensorFlow graph
+(`ssd_mobilenet_v2_coco_2018_03_29.pb`) and converted to ONNX for use with OpenCV's DNN module.
+
+## Model Details
+- **Architecture**: SSD (Single Shot MultiBox Detector) with MobileNet v2 backbone
+- **Input**: RGB image, 300×300, raw uint8, NHWC layout (`image_tensor:0`, shape `[1, 300, 300, 3]`)
+- **Output**: `detection_boxes:0` (normalized `ymin, xmin, ymax, xmax`), `detection_scores:0`, `detection_classes:0` (COCO class ids), `num_detections:0`
+- **Framework**: ONNX (converted from the TensorFlow frozen graph via tf2onnx, opset 18)
+- **Original weights**: http://download.tensorflow.org/models/object_detection/ssd_mobilenet_v2_coco_2018_03_29.tar.gz
+
+## Usage
+
+### Python
+```bash
+python demo.py --model ssd_mobilenet_v2_coco_2018_03_29_2026jul.onnx --image example_outputs/input_image.png --output example_outputs/output_image.png --conf 0.3
+```
+
+### C++
+The C++ demo runs inference with ONNX Runtime (C++ API) and uses OpenCV only for image I/O.
+Install ONNX Runtime (C++) from https://github.com/microsoft/onnxruntime/releases — this build
+uses `onnxruntime-linux-x64-1.25.0` — and adjust the ONNX Runtime and OpenCV paths to your setup:
+```bash
+ORT=/path/to/onnxruntime-linux-x64-1.25.0    # ONNX Runtime release dir (contains include/ and lib/)
+OCV=/path/to/opencv                          # OpenCV source tree
+OCVBUILD=/path/to/opencv/build               # OpenCV build directory (generated headers + libs)
+g++ -std=c++17 demo.cpp -o demo \
+  -I$ORT/include \
+  -I$OCV/include \
+  -I$OCV/modules/core/include \
+  -I$OCV/modules/imgproc/include \
+  -I$OCV/modules/imgcodecs/include \
+  -I$OCVBUILD \
+  -L$ORT/lib -Wl,-rpath,$ORT/lib -lonnxruntime \
+  -L$OCVBUILD/lib -Wl,-rpath,$OCVBUILD/lib -lopencv_imgcodecs -lopencv_imgproc -lopencv_core
+./demo --model ssd_mobilenet_v2_coco_2018_03_29_2026jul.onnx --image example_outputs/input_image.png --output example_outputs/output_image.png
+```
+
+## Conversion
+The ONNX model was exported from the frozen TensorFlow graph with tf2onnx (opset 18)
+via [convert_to_onnx.py](./convert_to_onnx.py) — input `image_tensor:0`, outputs
+`detection_boxes:0`, `detection_scores:0`, `detection_classes:0`, `num_detections:0`.
+Requires `tensorflow`, `tf2onnx`, and `onnx`.
+
+```bash
+python convert_to_onnx.py --pb ../pb/ssd_mobilenet_v2_coco_2018_03_29.pb
+```
+
+## License
+See [LICENSE](./LICENSE) — the model is released by the TensorFlow Authors under the Apache License 2.0.

ssd_mobilenet_v2_coco_2018_03_29/convert_to_onnx.py

@@ -0,0 +1,40 @@
+import argparse
+import datetime
+
+import onnx
+import tensorflow as tf
+import tf2onnx
+
+
+def load_graph_def(pb_path):
+    with tf.io.gfile.GFile(pb_path, "rb") as f:
+        graph_def = tf.compat.v1.GraphDef()
+        graph_def.ParseFromString(f.read())
+    return graph_def
+
+
+def main():
+    parser = argparse.ArgumentParser(description="Export ssd_mobilenet_v2_coco_2018_03_29.pb to ONNX")
+    parser.add_argument("--pb", default="../pb/ssd_mobilenet_v2_coco_2018_03_29.pb")
+    parser.add_argument("--opset", type=int, default=18)
+    args = parser.parse_args()
+
+    graph_def = load_graph_def(args.pb)
+
+    model_proto, _ = tf2onnx.convert.from_graph_def(
+        graph_def,
+        input_names=["image_tensor:0"],
+        output_names=["detection_boxes:0", "detection_scores:0", "detection_classes:0", "num_detections:0"],
+        opset=args.opset,
+    )
+    onnx.checker.check_model(model_proto)
+
+    stamp = datetime.datetime.now().strftime("%Y%b").lower()
+    onnx_path = "ssd_mobilenet_v2_coco_2018_03_29_%s.onnx" % stamp
+    with open(onnx_path, "wb") as f:
+        f.write(model_proto.SerializeToString())
+    print("wrote", onnx_path)
+
+
+if __name__ == "__main__":
+    main()

ssd_mobilenet_v2_coco_2018_03_29/demo.cpp

@@ -0,0 +1,98 @@
+#include <onnxruntime_cxx_api.h>
+#include <opencv2/imgproc.hpp>
+#include <opencv2/imgcodecs.hpp>
+#include <array>
+#include <cstdint>
+#include <iostream>
+#include <string>
+#include <vector>
+
+using namespace cv;
+
+static std::string argVal(int argc, char** argv, const std::string& key, const std::string& def)
+{
+    for (int i = 1; i + 1 < argc; ++i)
+        if (key == argv[i]) return argv[i + 1];
+    return def;
+}
+
+int main(int argc, char** argv)
+{
+    std::string model  = argVal(argc, argv, "--model",  "ssd_mobilenet_v2_coco_2018_03_29_2026jul.onnx");
+    std::string image  = argVal(argc, argv, "--image",  "example_outputs/input_image.png");
+    std::string output = argVal(argc, argv, "--output", "example_outputs/output_image.png");
+    float conf = std::stof(argVal(argc, argv, "--conf", "0.3"));
+
+    Mat img = imread(image);
+    if (img.empty())
+    {
+        std::cerr << "could not read image: " << image << std::endl;
+        return 1;
+    }
+
+    Mat rgb;
+    cvtColor(img, rgb, COLOR_BGR2RGB);
+    resize(rgb, rgb, Size(300, 300));
+    if (!rgb.isContinuous()) rgb = rgb.clone();
+
+    Ort::Env env(ORT_LOGGING_LEVEL_WARNING, "demo");
+    Ort::SessionOptions so;
+    Ort::Session session(env, model.c_str(), so);
+    Ort::AllocatorWithDefaultOptions alloc;
+
+    auto in_name = session.GetInputNameAllocated(0, alloc);
+    const char* in_names[] = {in_name.get()};
+
+    size_t out_count = session.GetOutputCount();
+    std::vector<Ort::AllocatedStringPtr> out_holders;
+    std::vector<std::string> out_str;
+    std::vector<const char*> out_names;
+    for (size_t i = 0; i < out_count; ++i)
+    {
+        out_holders.push_back(session.GetOutputNameAllocated(i, alloc));
+        out_str.push_back(out_holders.back().get());
+        out_names.push_back(out_str.back().c_str());
+    }
+
+    std::array<int64_t, 4> shape = {1, 300, 300, 3};
+    auto mem = Ort::MemoryInfo::CreateCpu(OrtArenaAllocator, OrtMemTypeDefault);
+    Ort::Value input = Ort::Value::CreateTensor<uint8_t>(mem, rgb.data, 300 * 300 * 3, shape.data(), shape.size());
+
+    auto outs = session.Run(Ort::RunOptions{nullptr}, in_names, &input, 1, out_names.data(), out_names.size());
+
+    const float *boxes = 0, *scores = 0, *classes = 0, *num = 0;
+    for (size_t i = 0; i < out_count; ++i)
+    {
+        const std::string& n = out_str[i];
+        if (n.find("detection_boxes") != std::string::npos) boxes = outs[i].GetTensorMutableData<float>();
+        else if (n.find("detection_scores") != std::string::npos) scores = outs[i].GetTensorMutableData<float>();
+        else if (n.find("detection_classes") != std::string::npos) classes = outs[i].GetTensorMutableData<float>();
+        else if (n.find("num_detections") != std::string::npos) num = outs[i].GetTensorMutableData<float>();
+    }
+    if (!boxes || !scores || !classes || !num)
+    {
+        std::cerr << "missing expected output tensors" << std::endl;
+        return 1;
+    }
+
+    int nd = (int)num[0];
+    int h = img.rows, w = img.cols;
+    Mat out = img.clone();
+    std::vector<std::string> lines;
+    for (int k = 0; k < nd; ++k)
+    {
+        if (scores[k] < conf) continue;
+        float ymin = boxes[k * 4 + 0], xmin = boxes[k * 4 + 1];
+        float ymax = boxes[k * 4 + 2], xmax = boxes[k * 4 + 3];
+        int cls = (int)classes[k];
+        rectangle(out, Point(int(xmin * w), int(ymin * h)), Point(int(xmax * w), int(ymax * h)), Scalar(0, 255, 0), 2);
+        putText(out, format("%d:%.2f", cls, scores[k]), Point(int(xmin * w), int(ymin * h) - 5),
+                FONT_HERSHEY_SIMPLEX, 0.5, Scalar(0, 255, 0), 1);
+        lines.push_back(format("%d %.3f %.3f %.3f %.3f %.3f", cls, scores[k], xmin, ymin, xmax, ymax));
+    }
+
+    imwrite(output, out);
+    std::cout << "ssd_mobilenet_v2_coco_2018_03_29 " << lines.size() << " detections" << std::endl;
+    for (const auto& l : lines) std::cout << l << std::endl;
+    return 0;
+}

ssd_mobilenet_v2_coco_2018_03_29/demo.py

@@ -0,0 +1,52 @@
+import argparse
+import glob
+import os
+
+import cv2 as cv
+import numpy as np
+import onnxruntime as ort
+
+here = os.path.dirname(os.path.abspath(__file__))
+
+
+def main():
+    parser = argparse.ArgumentParser(description="SSD MobileNet v2 COCO (ONNX) object detection demo")
+    parser.add_argument("--model", default=(glob.glob(os.path.join(here, "*.onnx")) or [""])[0])
+    parser.add_argument("--image", default=os.path.join(here, "example_outputs", "input_image.png"))
+    parser.add_argument("--output", default=os.path.join(here, "example_outputs", "output_image.png"))
+    parser.add_argument("--conf", type=float, default=0.3)
+    args = parser.parse_args()
+
+    img = cv.imread(args.image)
+    if img is None:
+        raise SystemExit("could not read image: %s" % args.image)
+
+    rgb = cv.resize(cv.cvtColor(img, cv.COLOR_BGR2RGB), (300, 300))
+
+    sess = ort.InferenceSession(args.model, providers=["CPUExecutionProvider"])
+    res = sess.run(None, {sess.get_inputs()[0].name: rgb[None].astype(np.uint8)})
+    onames = [o.name for o in sess.get_outputs()]
+    boxes = res[[i for i, n in enumerate(onames) if "detection_boxes" in n][0]].reshape(-1, 4)
+    scores = res[[i for i, n in enumerate(onames) if "detection_scores" in n][0]].reshape(-1)
+    classes = res[[i for i, n in enumerate(onames) if "detection_classes" in n][0]].reshape(-1)
+    nd = int(res[[i for i, n in enumerate(onames) if "num_detections" in n][0]].reshape(-1)[0])
+
+    h, w = img.shape[:2]
+    out = img.copy()
+    kept = []
+    for k in range(nd):
+        if scores[k] < args.conf:
+            continue
+        ymin, xmin, ymax, xmax = boxes[k]
+        kept.append((int(classes[k]), float(scores[k]), float(xmin), float(ymin), float(xmax), float(ymax)))
+        cv.rectangle(out, (int(xmin * w), int(ymin * h)), (int(xmax * w), int(ymax * h)), (0, 255, 0), 2)
+        cv.putText(out, "%d:%.2f" % (int(classes[k]), scores[k]), (int(xmin * w), int(ymin * h) - 5), cv.FONT_HERSHEY_SIMPLEX, 0.5, (0, 255, 0), 1)
+
+    cv.imwrite(args.output, out)
+    print("ssd_mobilenet_v2_coco_2018_03_29", len(kept), "detections")
+    for c, s, xmin, ymin, xmax, ymax in kept:
+        print(c, round(s, 3), round(xmin, 3), round(ymin, 3), round(xmax, 3), round(ymax, 3))
+
+
+if __name__ == "__main__":
+    main()

ssd_mobilenet_v2_coco_2018_03_29/ssd_mobilenet_v2_coco_2018_03_29_2026jul.onnx

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7ba2fdaa87b8cbbb52c16b5c6e31a7452c00e8ad68aec580bfb7b07f5b212619
+size 69584537