Closed NomiMalik0207 closed 5 years ago
Thank you @NomiMalik0207 For landmarks it's a 2 by 67 matrix corresponding to the (x, y) coordinates of 67 points of interests For the hog features I feed the vector directly to the cnn then I merge it with the landmarks and the dense layer after processing the images as shown in the diagram
great work man i want to know about how you are feeding landmarks and hog features to your network? either you are putting in the form of label of same like image?