This lecture overviews Transform Video Compression that has many applications in digital TV broadcasting, videoconferencing, video streaming and social media. It covers the following topics in detail: Spatiotemporal signal redundancy in video compression. Intraframe video coding (I-frames). Interframe video coding (P-, B-frames). Transform Video Coding (Block-based video coding, Discrete Cosine Transform, DCT Coefficient Quantization, Zig-zag… Continue reading Transform Video Compression
Archives: Resources
Resource description
Multiview Object Detection and Tracking
This lecture overviews Multiview Object Detection and Tracking that has many applications in video analysis and autonomous systems. It covers the following topics in detail: Multiview Human and Object Detection, Multiview Object tracking.
2D Object Detection and Tracking
This lecture overviews 2D Object Detection and Tracking that has many applications in autonomous car/drone vision (person, pedestrian, car tracking) and visual surveillance. It covers the following topics in detail: Object Detection and Forward Tracking. Object Forward-Backward Tracking.
2D Visual Object Tracking
Object/Target tracking is a crucial component of many vision systems. Object tracking issues are overviewed, e.g., occlusion handling, feature loss, drifting to the backgound. Many approaches regarding person/object detection and tracking in videos have been proposed. In this lecture, video tracking methods using correlation filters or convolutional neural networks are presented, focusing on video trackers… Continue reading 2D Visual Object Tracking
Motion Estimation
Motion estimation principals will be analyzed. Initiating form 2D and 3D motion models, displacement estimation as well as quality metrics for motion estimation will subsequently be detailed. One of the basic motion estimation techniques, namely block matching, will also be presented, along with three alternative, faster methods. A good overview of deep neural notion estimation… Continue reading Motion Estimation
Fast 3D Convolution algorithms
This lecture overviews Fast 3D Convolution algorithms that has many applications in the fast implementation of 3D image and video filtering, 3D CNNs and motion estimation. It covers the following topics in detail: 3D linear and cyclic convolutions, Fast 3D convolutions by using FFTs, Block-based methods, Optimal Winograd 3D convolutions.