Finding of Object Class in Visual Object Recognition

• Project/Area Number: 21800057
• Research Category: Grant-in-Aid for Research Activity Start-up
• Allocation Type: Single-year Grants
• Research Field: Intelligent informatics
• Research Institution: Tokyo Denki University
• Principal Investigator: HIDAKA Akinori Tokyo Denki University, 理工学部, 助教 (70553519)
• Project Period (FY): 2008 – 2010
• Project Status: Completed (Fiscal Year 2010)
• Budget Amount: ¥2,678,000 (Direct Cost: ¥2,060,000、Indirect Cost: ¥618,000); Fiscal Year 2010: ¥1,274,000 (Direct Cost: ¥980,000、Indirect Cost: ¥294,000); Fiscal Year 2009: ¥1,404,000 (Direct Cost: ¥1,080,000、Indirect Cost: ¥324,000)
• Keywords: パターン認識 / 機械学習 / 画像認識 / クラスタリング / 人工知能 / 知能機械

Research Abstract

In visual object recognition based on pattern recognition, unsupervised learning is a still important problem. In this research, we try semi-automatic generating of training samples from non-labeled video sequence by using unsupervised approach. We use grow cut and template matching based object traking method for our sample extraction. In our experiments, we show that our approach can make the clusters of targets without teacher signals.