NORJ commited on
Commit
c4fa27e
Β·
verified Β·
1 Parent(s): 66a2958

Upload 4 files

Browse files
Files changed (4) hide show
  1. README.md +41 -20
  2. best_model.h5 +3 -0
  3. local_finetuned_model.h5 +3 -0
  4. malaria_model.tflite +3 -0
README.md CHANGED
@@ -1,20 +1,41 @@
1
- ---
2
- license: mit
3
- language:
4
- - en
5
- metrics:
6
- - accuracy
7
- - precision
8
- - recall
9
- - f1
10
- base_model:
11
- - google/mobilenet_v2_1.0_224
12
- pipeline_tag: image-classification
13
- tags:
14
- - Malaria
15
- - Deep-learning
16
- - Healthcare
17
- - Medical-imaging
18
- - Cell-classification
19
- - Ghana
20
- ---
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # AI Malaria Detection System - Ghana Model
2
+
3
+ ## Overview
4
+ This system uses deep learning (MobileNetV2) to detect malaria parasites in blood smear images. The model has been fine-tuned specifically on Ghanaian blood smear samples to ensure accuracy with local staining and imaging conditions.
5
+
6
+ ## Performance Metrics
7
+ - **Overall Accuracy**: 97.2%
8
+ - **Parasitized Detection**: 97.8%
9
+ - **Uninfected Rejection**: 92.3%
10
+ - **Model Size**: 2.54 MB (mobile-optimized)
11
+
12
+ ## Folder Structure
13
+ ```
14
+ /malaria_detector/
15
+ β”œβ”€β”€ models/
16
+ β”‚ β”œβ”€β”€ local_finetuned_model.h5
17
+ β”‚ └── malaria_model.tflite
18
+ β”œβ”€β”€ scripts/
19
+ β”‚ β”œβ”€β”€ predict_patches.py
20
+ β”‚ └── requirements.txt
21
+ β”œβ”€β”€ evaluation/
22
+ β”‚ β”œβ”€β”€ final_report.pdf
23
+ β”‚ └── confusion_matrix.png
24
+ └── README.md
25
+ ```
26
+
27
+ ## Deployment & Usage
28
+ 1. **Clone the Repository**: Download the entire `malaria_detector` folder.
29
+ 2. **Install Dependencies**: Navigate to the `scripts/` folder and install required libraries:
30
+ ```bash
31
+ pip install -r requirements.txt
32
+ ```
33
+ 3. **Model Location**: Ensure `local_finetuned_model.h5` and `malaria_model.tflite` are in the `models/` folder.
34
+ 4. **Run Prediction**: Use the `predict_patches.py` script to analyze whole slide images:
35
+ ```bash
36
+ python scripts/predict_patches.py /path/to/your/image.jpg [confidence_threshold]
37
+ ```
38
+ * `[confidence_threshold]` is optional; defaults to `0.5` (50%). Recommended: `0.5`.
39
+
40
+ ## Contact
41
+ For questions or further assistance, please contact the development team.
best_model.h5 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:78a322f1814d483654abca7fe053aef096e8d41a0c614fbc266fb9532c7cb9f2
3
+ size 11540648
local_finetuned_model.h5 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:177f02f505c4384b1dd11a44ac66c1682357e4aa3e448e0555ba6afcc6dd974d
3
+ size 17305376
malaria_model.tflite ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:fe24b179408842b3647ad02dfecdc491db070e1d2fc5cb77ddad5342b3b46c1e
3
+ size 2667336