andyzeng MIT-Princeton Vision Toolbox for the Amazon Picking Challenge 2016 - RGB-D ConvNet-based object segmentation and 6D o…