File size: 2,019 Bytes
6f3ad83 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 | """
SignLens - Inference Example
Demonstrates how to load and run inference on the SignLens ISL & ASL models.
"""
import pickle
import numpy as np
def load_isl_model(model_path="isl_landmark_model.pkl"):
"""Load the 156-D ISL Dual-Hand Landmark Classifier."""
with open(model_path, "rb") as f:
data = pickle.load(f)
print(f"Loaded ISL Model: {len(data['classes'])} classes, {data['num_features']}-D input")
return data["model"], data["classes"]
def load_asl_model(model_path="asl_landmark_model.pkl"):
"""Load the 78-D ASL Single-Hand Landmark Classifier."""
with open(model_path, "rb") as f:
data = pickle.load(f)
print(f"Loaded ASL Model: {len(data['classes'])} classes, {data['num_features']}-D input")
return data["model"], data["classes"]
def predict_sign(model, classes, features):
"""
Predict sign class and confidence score.
:param model: Trained sklearn MLPClassifier
:param classes: List of class labels
:param features: 1D array of 78 (ASL) or 156 (ISL) normalized landmark features
:return: (predicted_label, confidence_score, top3_predictions)
"""
probs = model.predict_proba([features])[0]
top_idx = int(np.argmax(probs))
top_label = classes[top_idx]
top_conf = float(probs[top_idx])
top3_idx = np.argsort(probs)[::-1][:3]
top3 = [(classes[i], float(probs[i])) for i in top3_idx]
return top_label, top_conf, top3
if __name__ == "__main__":
# Test ISL Model
isl_model, isl_classes = load_isl_model()
dummy_isl_features = np.zeros(156, dtype=np.float32)
label, conf, top3 = predict_sign(isl_model, isl_classes, dummy_isl_features)
print(f"[ISL Test] Dummy Input Prediction: {label} ({conf:.2%})")
# Test ASL Model
asl_model, asl_classes = load_asl_model()
dummy_asl_features = np.zeros(78, dtype=np.float32)
label, conf, top3 = predict_sign(asl_model, asl_classes, dummy_asl_features)
print(f"[ASL Test] Dummy Input Prediction: {label} ({conf:.2%})")
|