Tutorial: Local Feature Extraction and Learning for Computer Vision HD

Introduction and Brief Review of Classical Feature Descriptors, Pascal Fua (EPFL) Modern Descriptors: Towards High Matching Performance, Bin Fan (Chinese Academy of Sciences) Modern Descriptors: Towards High Matching Efficiency, Jiwen Lu (Tsinghua Univ.) Summary & Discussion, Pascal Fua (EPFL), Bin Fan (Chinese Academy of Sciences), Jiwen Lu (Tsinghua Uni.) Local feature is at the core of many computer vision tasks and has played an important role in the development of various computer vision applications. A large number of local features oriented for different tasks have been proposed over the past two decades by using either hand-crafted methods or learning-based techniques. This tutorial will give an extensive introduction of the latest advances on this topic by dividing them into two categories: those were proposed for high matching accuracy, and those were introduced for high efficiency performance. Typical computer vision applications based on these introduced local features will also be discussed.