face-alignment-mtcnn
face-alignment-mtcnn copied to clipboard
TypeError when using align() on certain images in facealignment library
Description I'm encountering a TypeError when attempting to align certain images using the facealignment library's align() function. While the function works as expected with some images, it throws an error with others, despite both types of images being seemingly similar (e.g., both are 3-channel images). import facealignment det2 = facealignment.FaceAlignmentTools() aligned_img, b, points= det2.align(img) # returning boxes, n_src_points from align function The function is expected to return the aligned image, bounding box, and facial landmarks without any errors, as it does with other images.
TypeError Traceback (most recent call last) Cell In[40], line 2 1 bboxes1, pps1, ccs1 = det1.call(img) ----> 2 aligned_img, b, points= det2.align(img) 3 # screen_img = cv2.hconcat([img, aligned_img]) 6 pps1
File c:\Users\Mary\Documents\ais-projects-software_ai_features-Cheetah_B0_Receiver-07_Software\ais-projects-software_ai_features-Cheetah_B0_Receiver-07_Software\Cheetah_B0_Receiver\07_Software\facealignment\tools.py:202, in FaceAlignmentTools.align(self, img, dsize, allow_multiface, central_face, alignment) 199 dst_points = self._landmarks[self._alignment_style] # Load saved landmarks 201 if alignment is None: --> 202 n_src_points, boxes= self.detect_face(img, allow_multiface=allow_multiface or central_face) 203 else: 204 n_src_points = alignment
TypeError: cannot unpack non-iterable NoneType object
Environment Python version: 3.9.18 Operating System: windows10
Additional Context Both images that I tested (the one that works and the one that doesn't) are 3-channel RGB images and appear correctly when displayed. This suggests the issue may not be with the image format or structure.