Professor William Freeman studied computer vision for his PhD in 1992 from MIT. He later received his BS in physics and MS in electrical engineering from Stanford in 1979, and an MS in applied physics from Cornell in 1981. From 1981-1987, Freeman worked at the Polaroid Corporation as the co-developer of an electronic printer (Polaroid Palette) and developed algorithms for color image reconstruction used in Polaroid’s electronic camera. Now, Freeman is a professor in the electrical engineering and computer science department at MIT. As well as a member of the computer science and artificial intelligence laboratory (CSAIL) since 2001. Some of his awards include: Outstanding Paper prize at the Conference on Computer Vision and Pattern Recognition (1997), member of the IEEE PAMI TC Awards Committee, and associate editor of IEEE Trans on Pattern Analysis and Machine Intelligence (IEEE-PAMI).
Industry Impact
William's current research interests include machine learning applied to computer vision, Bayesian models of visual perception, and computational photography.
Visual Dynamics: Probabilistic Future Frame Synthesis via Cross Convolutional Networks
This project studies the problem of synthesizing a number of likely future frames from a single input image. A probabilistic model makes it possible to sample and synthesize many possible future frames from a single input image. Future frame synthesis can be challenging since it involves low and high-level image and motion understanding. We propose a novel network structure, namely a Cross Convolutional Network to aid in synthesizing future frames; this network structure encodes image and motion information as feature maps and convolutional kernels. This model performs well on synthetic data like 2D shapes, animated game sprites, and real-world videos. In addition, it can be applied to tasks such as visual analogy-making, and present an analysis of the learned network representations.
3D Generative Adversarial Networks
In this research study, we propose a novel framework 3D Generative Adversarial Network (3D-GAN), which generates 3D objects from a probabilistic space by leveraging recent advances in volumetric convolutional networks and generative adversarial nets. The benefits of our model (1) use adversarial criterion and enables the generator to capture object structure implicitly and to synthesize high-quality 3D objects (2) sample objects without a reference image or CAD models and explore the 3D object manifold and (3) the adversarial discriminator provides a powerful 3D shape descriptor which has 3D object recognition.
Learning the Arrow of Time from Videos
Most fundamental equations look the same whether time is playing forward or backwards. The collision of two particles look the same if time is reversed. Yet at a macroscopic scale, time is not reversible and the world behaves differently whether time is playing forwards or backwards. In this study, the aim is to train machines to visually detect the arrow of time, if a video is playing forward or backward, by watching millions of online clips.