Abstract

This chapter explores some hypothetical computer vision pipeline designs to understand HW/SW design alternatives and optimizations. Instead of looking at isolated computer vision algorithms, this chapter ties together many concepts into complete vision pipelines. Vision pipelines are sketched out for a few example applications to illustrate the use of different methods. Example applications include object recognition using shape and color for automobiles, face detection and emotion detection using local features, image classification using global features, and augmented reality. The examples have been chosen to illustrate the use of different families of feature description metrics within the Vision Metrics Taxonomy presented in Chap. 5. Alternative optimizations at each stage of the vision pipeline are explored. For example, we consider which vision algorithms run better on a CPU versus a GPU, and discuss how data transfer time between compute units and memory affects performance. Document type: Part of book or chapter of book

Abstract

Action recognition in videos is a challenging task due to the complexity of the spatio-temporal patterns to model and the difficulty to acquire and learn on large quantities of video data. Deep learning, although a breakthrough for image classification and showing promise for videos, has still not clearly superseded action recognition methods using hand-crafted features, even when training on massive datasets. In this paper, we introduce hybrid video classification architectures based on carefully designed unsupervised representations of hand-crafted spatio-temporal features classified by supervised deep networks. As we show in our experiments on five popular benchmarks for action recognition, our hybrid model combines the best of both worlds: it is data efficient (trained on 150 to 10000 short clips) and yet improves significantly on the state of the art, including recent deep models trained on millions of manually labelled images and videos. Document type: Part of book or chapter of book