Shu Kong

Hello, thanks for your interest in trying the demo. The mat files containing GT are a format I used for training and evaluation at my own side. I didn't put...

Hi, Ahyun, Thanks for your interest! I'm not going to release the post-processing&evaluation code including ranking. Basically what I did is to collect morphological features of the proposal candidate, like...

Hi, Ahyun, Yes, you are right. I build a voting method to combine predicted class labels and the instance proposals. The idea is basically about counting the major class labels...

Yes. You can think in this way. The mean-shift behaves more like a loss rather than parametric module. But here are two things to notice -- 1. In the toy...

Hello, thank you for your interest! If you are using matlab, you can simply type "*save('filename.mat', 'var1_to_save', 'var2_to_save')*" Here is the official link about "save" https://www.mathworks.com/help/matlab/ref/save.html If you are using...

Hi, if you have full body dataset, you can definitely train such models for that. It must be interesting to compare a simple feedforward model with the bilinear version. On...

It makes sense to do gender classification over whole body especially when the face is occluded or blurry. This should require you make the correct label (female&male) to enable good...

No, there is no caffe version of the low-rank bilinear method. There are several reasons in providing demo 1 -- 1. Caffe runs a little faster than matconvnet, so simply...

Hi, Thanks for trying this out. This happens at my end also. The reason I guess is that updating the big kernel from PCA is non-trivial. I still don't have...

Hi, For DTD dataset, I trained the models progressively -- training the straightforward baseline with cross-entropy loss, then fine-tuning for bilinear CNN, and finally fine-tuning for LRBP. As you are...