Open BraaZain opened 5 years ago
Sure, but what did you already try? It should just involve combining the KNN example with the facerec from video example.
i only changed the main to take a frame from video capture and send it to the predict function instead of the image path :
if __name__ == "__main__":
#step1: use open cv to create a video capture
video_capture = cv2.VideoCapture(0)
while True:
ret, frame = video_capture.read()
rgb_frame = cv2.cvtColor(frame, cv2.COLOR_BGR2RGB)
#get predictions
predictions = predict(x_img = rgb_frame, model_path = "Classifier.plk")
#show predictions
result_img = show_predictions(rgb_frame, predictions)
#display the video
final_frame = np.array(result_img)
final_frame = cv2.cvtColor(final_frame, cv2.COLOR_RGB2BGR)
cv2.imshow("test", final_frame)
if cv2.waitKey(1) & 0xFF == ord('q'):
break
video_capture.release()
cv2.destroyAllWindows()
it worked but the video is really slow , i don't know if it is because my laptop specification or the code is slow?? could you suggest me something to fix this??
is there any way to use this KNN Classifier with open cv to predict faces in a live video. i have tried to combine them but didn't find a way to do it.