Beckman Team Members Score Highest at Image Competition

Beckman Institute researcher Tom Huang led a team that finished with the top three scores at the ImageNet Large Scale Visual Recognition Challenge 2010. The competition tests teams from around the world on the accuracy of their object classifier algorithms used to retrieve and identify content in images.

Beckman Institute researcher Tom Huang led a team that finished with the top three scores at the ImageNet Large Scale Visual Recognition Challenge 2010. The competition tests teams from around the world on the accuracy of their object classifier algorithms used to retrieve and identify content in images.

The team was composed of Huang, and members of the Image Formation and Processing (IFP) group he heads at Beckman, as well as members from the NEC Corporation and a member from Rutgers University. IFP team members are Liangliang Cao, Zhen Li, Min-Hsuan Tsai, and Xi Zhou. 

The competition was held in conjunction with Visual Object Classes Challenge 2010 sponsored by the Pattern Analysis, Statistical Modeling and Computational Learning (PASCAL) Network of Excellence. ImageNet is an ongoing research effort to provide researchers around the world an easily accessible image database.

The goal of the competition required teams to estimate the content of photographs for the purpose of retrieval and automatic annotation using a subset of the large hand-labeled ImageNet dataset (10,000,000 labeled images depicting 10,000+ object categories) as training. Each team’s algorithms had to produce labelings specifying what objects are present in the images. The team’s algorithms were judged based on their ability to classify objects correctly.

The Illinois-NEC team had the top three highest scores with their top score scoring a 72 percent accuracy rate, more than five percent better than their nearest team in the competition.