A Holistic Approach to Robust and Scalable Object Recognition

This project is conducted at Advanced Digital Sciences Center (ADSC), as part of the Interactive Digital Media (IDM) sub-program. The PI leading our research efforts is Dr. Yi Ma from UIUC and MSRA.


[Left plot] Training acquisition system: Four projectors and two cameras controlled by one computer. [Top-right plot] 38 traning images of one subject captured by the training acquisition system. [Bottom-right plot] Representative testing examples, including indoor/outdoor images (1st row), images of subjects with eyeglasses (2nd row), and image of subjects with sunglasses (3rd row).


One of Singapore’s strategic plans is to become the center of innovation for the next generation information technologies. ICM is playing an increasingly crucial role in the future IT sector, largely due to its vast potential applications in numerous commercial products and services. In recent days, computer vision techniques such as face recognition, image search have started to generate impact in many commercial products. IT giants such as Google, Microsoft, and Apple etc have significantly increased their investment in this area. The holistic approach proposed in this project represents an entirely new approach for face and object recognition in general. By leveraging some of the most powerful modern optimization techniques, it allows us to harness all information in the entire images, saving tremendous trouble by eliminating the unstable step of local feature extraction. It has been marked as the core technology for next generation high-quality face recognition system. Singapore government traditionally has strong interests in security technologies like face recognition and license plate recognition. An important goal of this project is for ADSC and A*STAR to forge a union and try to bring this emerging new technology to its next level of maturity. Certainly the commercial impact of such technology is not going to be limited to the Singapore market as face recognition technologies are also in high demand in many other Asian countries like China and Japan, or even back to the US. In addition, through this project, we will be able to explore the use of such new techniques for a broader range of object recognition tasks far beyond face recognition, which will help generate even more exciting opportunities for Singapore.