With the advances in GPU’s computing power, it is now feasible to train very deep neural networks to solve problems which have not been made possible before. In the past 5 years, we have witnessed the success of applying deep learning to many computer vision problems. Novel deep learning based methods have been proposed for solving various vision tasks, and they have demonstrated unprecedented robustness and accuracy in their results. In this workshop, we will introduce some representative methods for a number of vision tasks, including image classification, object detection, scene reconstruction, and image synthesis. It is hoped that this short introductory talk could serve as a starting point for those who would like to employ computer vision techniques in solving their problems, and for those who would like to engage in computer vision research.