机器学习大牛主页，这个分类比较杂

1. piotrdollár，
   研究方向： Object Detection，
   Pose Estimation，
   Caltech Pedestrian Benchmark，
   Multiple Instance & Multiple Component Learning，
   Non-Isometric Manifold Learning ，
   Boundary & Feature Learning ，
   Behavior Recognition & Animal Behavior

2. Ma Yi
   Low-Rank Matrix Recovery via Convex Optimization.
   Face Recognition via Sparse Representation
   Clustering and Classification via Lossy Compression
   Generalized Principal Component Analysis
   Multiple-View Geometry and 3D Reconstruction

3. SVETLANA LAZEBNIK
   Images and Text，Image Parsing，Scene Understanding，Similarity- Preserving Binary Codes，Landmark Photo Collections，Texture and Object Recognition Using Local Features等等