Category Archives: International

Robust Online Multi-Object Tracking based on Tracklet Confidence and Online Discriminative Appearance Learning– Author :  Seung-Hwan Bae and Kuk-Jin Yoon – Published Date : June 24, 2014 – Category : Tracking and Detection – Place of publication : IEEE International Conference on Computer Vision and Pattern Recognition (CVPR) Abstract Online multi-object tracking aims at producing complete tracks of multiple objects…

Robust Online Multi-Object Tracking with Data Association and Track Management– Author :  Seung-Hwan Bae and Kuk-Jin Yoon – Published Date : July, 2014 – Category : Tracking and Detection – Place of publication : IEEE Transactions on Image Processing (TIP) Abstract In this paper, we consider a multi-object tracking problem in complex scenes. Unlike batch tracking systems using detections of…

Polyp Detection via Imbalanced Learning and Discriminative Feature Learning– Author: Seung-Hwan Bae and Kuk-Jin Yoon – Published Date: May 18, 2015 – Category: Tracking and Detection – Place of publication: IEEE Transactions on Medical Imaging Abstract Recent achievement of the learning based classification leads to the noticeable performance improvement in automatic polyp detection. Here, building large good datasets is very crucial for learning a reliable detector. However, it is practically challenging due to the diversity of polyp types, expensive inspection, and labor- ntensive labeling tasks. For this reason, the polyp datasets usually tend to be imbalanced, i.e. the number of non polyp samples is much larger than that of polyp samples, and learning with those imbalanced datasets results in a detector biased toward a non polyp class. In this paper, we propose a data sampling based boosting frameworkto learn an unbiased polyp detector from the imbalanced datasets. In our learning…

Polyp Detection via Imbalanced Learning and Discriminative Feature Learning– Author: Seung-Hwan Bae and Kuk-Jin Yoon – Published Date: May 18, 2015 – Category: Tracking and Detection – Place of publication: IEEE Transactions on Medical Imaging Abstract Recent achievement of the learning based classification leads to the noticeable performance improvement in automatic polyp detection. Here, building large good datasets is very crucial for learning a reliable detector. However, it is practically challenging due to the diversity of polyp types, expensive inspection, and labor- ntensive labeling tasks. For this reason, the polyp datasets usually tend to be imbalanced, i.e. the number of non polyp samples is much larger than that of polyp samples, and learning with those imbalanced datasets results in a detector biased toward a non polyp class. In this paper, we propose a data sampling based boosting frameworkto learn an unbiased polyp detector from the imbalanced datasets. In our learning…