Coursera Flash Sale
40% Off Coursera Plus for 3 Months!
Grab it
Explore cutting-edge developments in computer vision through this advanced graduate-level course that delves deep into transformer architectures and their revolutionary applications in visual understanding. Master the fundamentals of artificial intelligence and deep learning before progressing to sophisticated transformer models including Vision Transformers (ViT), Data-efficient Image Transformers (DeiT), and Swin Transformers with their hierarchical windowing approaches. Investigate how transformers have transformed object detection through DETR and Deformable DETR architectures, enabling end-to-end detection without traditional anchor-based methods. Examine video understanding capabilities through Video Vision Transformers (ViViT), Video Swin Transformers, and space-time attention mechanisms that capture temporal dynamics in visual sequences. Discover emerging applications in multimodal learning, including CLIP's approach to learning transferable visual models from natural language supervision and zero-shot text-to-image generation techniques. Analyze advanced topics such as high-resolution image synthesis using taming transformers, semantic segmentation from sequence-to-sequence perspectives, and the intriguing properties that emerge in self-supervised vision transformers. Study general perception frameworks like Perceiver that use iterative attention mechanisms and explore how ConvNets are being reimagined for modern computer vision challenges. Engage with contemporary research through student project presentations that demonstrate practical applications of these advanced concepts, providing hands-on experience with state-of-the-art computer vision methodologies and transformer-based architectures.
