I choose to work on the problem of sketch-based retrieval for this semester, where the goal is to find relevant 3D models in a large collection. The input is a rough 2D sketch, rather than a complete 3D model. This problem is challenging and will alows me to use very well-developed techniques from image analysis. My goal will be to read and implement a recent paper, Sketch-Based Shape Retrieval, and test it on any dataset, such as the one described here.