rec_vis:
	Contains the images rendered by running a model pre-trained on ModelNet.
	For example, rec_res-airplane_0630-point.png means this is an image corresponding to the mesh in the category airplane, and its identification in the category is 0630.
seg_vis:
	Contains the images rendered from semi-supervised training on ShapeNetPart Segmentation dataset.
	for example, results-02691156-4cee36a2e8dd3b24b87697d3904b168b-gt.png means this is a ground truth image from category 02691156, and its identification in the category is
4cee36a2e8dd3b24b87697d3904b168b. if gt is replaced with pred, then it is a rendered image for prediction.