The process of evolution accelerated on the planet earth 543 million years ago. For the evolution to suddenly accelerate something extraordinary happened, the organism started developing a vision. Now rather than waiting for food, the organisms could see food and move towards them and the food could see the predator coming towards them and run away to save themselves. This is how vision began, but I am not here to tell the story how the vision developed in animals. I am here to talk about the computer vision which is inspired by the natural vision system. When I say vision it is not just ability to see but also ability to understand.
In summer of 1966, computer vision started as a summer project in MIT AI lab. It has been 51 years and we are yet to solve the problem. The picture below is the memo written on July 7, 1966, to start the project of computer vision as a summer project.
One other interesting fact. The AI lab was established before the computer department. It shows that computer science is not just computers rather it is an approach in solving a problem. As a student of computer science, I feel like people do understand or ignore this fact.
The computer vision is a replica of biological vision system which consists of eyes that can see and primary vision cortex in the brain that process the information. The insight of the biological vision system was given by Hubel and Wiesel at Harvard.They performed an experiment to learn about primary vision cortex. They showed pictures to a sedated but conscious cat and recorded its response. They found out that neurons in the primary vision cortex were arranged in series of columns and each column is only interested in some small feature of the image.
In 1982, David Marr published a book called Vision. In this book, he wrote the 3-D representation is hierarchical. The image below shows an abstract vision system.
Based on these insights the computer vision has developed to extend such that facebook can tell what are different objects and who are the people in the image. But we still have a long way to go.