Computer Vision Applications - Full Stack Deep Learning - Spring 2021

Explore notable applications of deep learning in computer vision in this 43-minute lecture. Dive into various ConvNet architectures, including AlexNet, ZFNet, VGGNet, GoogLeNet, ResNet, and SqueezeNet. Examine localization, detection, and segmentation problems, along with methods like Overfeat, YOLO, and SSD. Investigate region proposal techniques such as R-CNN, Faster R-CNN, Mask R-CNN, and U-Net. Discover advanced tasks like 3D shape inference, face landmark recognition, and pose estimation. Conclude with insights on adversarial attacks and style transfer, providing a comprehensive overview of cutting-edge computer vision applications.

- Introduction
- AlexNet
- ZFNet
- VGGNet
- GoogLeNet
- ResNet
- SqueezeNet
- Architecture Comparisons
- Localization, Detection, and Segmentation Tasks
- Overfeat, YOLO, and SSD Methods
- Region Proposal Methods R-CNN, Faster R-CNN, Mask R-CNN, U-Net
- Advanced Tasks 3D Shape Inference, Face Landmark Recognition, and Pose Estimation
- Adversarial Attacks
- Style Transfer