PASADENA, Calif. – 2 Feb 11
UC San Diego Student Researches Intelligent Computer Vision
By: Heather L. Ogletree
The future is here.Motion pictures and even cartoons from the past envisioned intelligent homes and cars, flat screen televisions that could be used for communication, and humanoid robots. Many of these things are now in existence (smart cars, the iPad, Robonaut) and others are in the works through the space age research conducted at NASA. At the Jet Propulsion Laboratory (JPL) in Pasadena, Calif., researchers are working to create something just as futuristic: Intelligent Computer Vision.
According to Dr. Thomas Lu, Senior Researcher of the Optical Image Processing Team in Bio-Inspired Technologies & Systems Group at JPL, “The research is directly applicable to autonomous guidance of spacecraft landing, docking, rendezvous and hazardous avoidance.... The autonomous target recognition system helps the robots and autonomous vehicles to understand the environment, and perform autonomous maneuvers. It would also help next generation computers to understand the content of digital images and videos in this digital age.”
This fall, Dr. Lu chose Hank Lin from the Undergraduate Student Research Program (USRP) to act as a Research Assistant on his project concerning the field of Intelligent Computer Vision. Lin, a naturalized U.S. citizen from Taipei, Taiwan, had already participated in the USRP internship experience in 2009, so when the University of California, San Diego student arrived, he had high hopes for his time at NASA.
As part of Lin’s project, he helped “train intelligent computer programs to automatically detect, recognize and track objects from CCD/CMOS imaging cameras.”
Lin began his project at NASA during the end stages of the research and development, which called for a high learning curve. He stated, “When I reached NASA, I [did] not know computer vision at all.” However, he indicated that his background in machine learning, math and his past research experience helped in overcoming this challenge. In 2007, Lin completed a year abroad at Japan’s Osaka University where he taught himself JAVA and worked on an independent study project concerning artificial intelligence and machine learning.
Through his USRP project, Lin said, “I was able to learn and integrate different machine learning algorithms: neural network, adaboost, and SVM [Support Vector Machine].... You work with very smart people, and it just drives everyone to excel.” Since interning, Lin has co-authored two scientific papers, and these papers were presented in the International SPIE Defense, Security and Sensing Conferences.
After college, Lin plans to attend graduate school. He said, “I feel there is much learning and training needed for me to do better work.”
As for NASA, Lin said, “When people ask me, ‘What do you want to become,’ I always answer, ‘My dream is to work at NASA, to send robots into space." And working with USRP strengthened that passion. “You know what you do will be used on a mission, and it felt great to have a small contribution to a big plan,” concluded Lin.