Ahmednull / L2CS-Net

The official PyTorch implementation of L2CS-Net for gaze estimation and tracking
MIT License
341 stars 81 forks source link

ValueError Attempting to Process Frames Without Faces #22

Open kbvatral opened 1 year ago

kbvatral commented 1 year ago

Running into a numpy ValueError: need at least one array to stack when running the pipeline on a frame of video that has no faces detected. Seems to point back to line 90 in the Pipeline file: pitch, yaw = self.predict_gaze(np.stack(face_imgs)). Using Numpy 1.24.3

My guess is that this is due to face_imgs being an empty list when there are no detected faces. Working on a fix in my fork and will PR when complete.