In this assignment, your task is to implement a program that propagates sparse labels to among an entrie shape collection.
You program will be tested on this shape collections : Data
1. Build a shape graph among the input shapes. This could be done by comparing shape descriptors (e.g., D2 and light field).
2. Convert the problem as solving a multi-label MRF problem, where the states associated with each shape are potential labels. Unary terms may encode the label instances, and binary terms may encode the similarity graph.
Summarize your findings in a single report.pdf. The report should include visualizations of the classified shapes.