Embark on an intellectual journey through the evolution of computer vision with renowned expert Keechin Goh at SuperAI 2024. This compelling talk, "A Primer to Modern Day Computer Vision," takes you from the foundational concepts of the field to its groundbreaking applications today.

​

Keechin Goh explores the historical roots of computer vision, from its nascent stages in the 1970s to the major developments that have empowered the technology we use today. Understand why computer vision is uniquely complex due to its unstructured nature, compared to structured data types like financial spreadsheets or language models. Discover how early methods such as edge detection paved the way for more advanced technologies like scale-invariant feature transform (SIFT) and convolutional neural networks (CNN), and learn why CNNs remain invaluable despite the industry's transition to newer methodologies.

​

Goh delves into the innovations of vision transformers (ViTs), highlighting their ability to fuse traditional image processing techniques with modern transformer architectures to enhance scene understanding. Gain insights into zero-shot learning and vector search, empowering systems to operate with minimal or indirect training, a testament to the strides made in reducing computational and memory demands.

​

Stay at the forefront of technological advancement as you explore recent developments in multi-modal AI models like Google's PaLM 2, and discover the practical implications of these advances in fields like autonomous driving and niche manufacturing. As competitive vision continues to evolve, learn about privacy-preserving techniques and how companies are leveraging proprietary data for bespoke applications.

​

Join the conversation on the future of computer vision. Like, comment, and subscribe to stay updated with more enriching content from SuperAI 2024.