박사과정 배성호 논문 - ICIP2014 국제학술대회 발표 승인
Title: Object tracking based on online Partial Instance Learning with multiple local strong classifiers
Many conventional appearance models with online boosting (OB) frameworks for object tracking utilize one single strong classifier for the global region of an image example. However, this single strong classifier is often unstably updated towards a sub-optimal solution when partial occlusions or changes of an appearance occur. In order to overcome this problem, we propose a new appearance model based on Partial Instance Learning (PIL) with multiple local strong classifiers. The key idea of PIL is that image examples are divided into several partial image examples (or local-images), each of which is then independently trained with a local strong classifier. Finally, a tracker is updated for the optimal solution in the sense that the joint probability of partial image examples for each input image example becomes the largest. The proposed PIL method can be considered a risk diversification strategy for unpredictable partial occlusions or appearance changes of an object. Also, it can be regarded as a divide-and-conquer method of OB, so that PIL only requires approximately 20% of computations compared with other OB methods in terms of iterations taken for learning process. Experiment results show that the proposed PIL-based object tracking method achieves better performance in tracking accuracy and much faster processing speed than other compared real-time based object tracking methods.