Did you the depth cues,can you privide preprocessing the depth file?

[RichardTaoK]

Hello, I'd like to use the network vgg16_deep_model，but i found in model.py ,the line 26,and 28,deep=images[:,:,:,3:4],but,images.shape=2242243, i check the preprocess.py,the image.shape still is 2242243,so ,if you use the depth cue,both the dataset has no depth image,so ,where did you get the depth cue,if you use ,can you chat with me

[way151]

Hello, I have a similar problem. The model in the corresponding paper has 'Depth-enhanced Subsidiary Network',i.e., 'We ... incorporate the depth-based features acquired by the subsidiary
network as a convolution layer into the proposed master network', but I didn't find it in model.py . I'm sorry but do I miss sth?