https://scholars.lib.ntu.edu.tw/handle/123456789/581451
標題: | How Incompletely Segmented Information Affects Multi-Object Tracking and Segmentation (MOTS) | 作者: | Chou Y.-S Wang C.-Y SHOU-DE LIN Liao H.-Y.M. |
關鍵字: | Deep learning; Image segmentation; Object detection; Object recognition; Pixels; Semantics; Baseline models; Bounding box; Multi-object tracking; Object detectors; Semantic segmentation; State of the art; Tracking by detections; Tracking performance; Object tracking | 公開日期: | 2020 | 卷: | 2020-October | 起(迄)頁: | 2086-2090 | 來源出版物: | Proceedings - International Conference on Image Processing, ICIP | 摘要: | In recent years, deep learning has made dramatic advances in computer vision field, especially in improving the performance of object detection as well as instance semantic segmentation. Still, multi-object tracking (MOT) remains a very challenging issue. Even in state-of-the-art deep learning-based object detectors, a preferred paradigm for MOT: tracking-by-detection, can only slightly improve the tracking performance. Pixel-level information is considered more precise and useful for tracking performance improvement than using conventional information, such as foreground or background content in a bounding box. However, the performance of current state-of-the-art models for automatically annotating pixel-level information is still far from the expectation of human beings. Therefore, we shall explore how multi-object tracking and segmentation (MOTS) is affected when the information obtained after applying instance semantic segmentation is incomplete. We propose a mask-guided two-streamed augmentation learning (MGTSAL) algorithm, which can be applied to TrackR-CNN to alleviate significant drop of MOTS performance when encountering incompletely segmented information. We evaluate the proposed approach on MOTS KITTI dataset, and our approach outperforms the baseline model TrackR-CNN in all our experimental settings. The promising experimental results and ablation study validate the effectiveness of the proposed approach. ? 2020 IEEE. |
URI: | https://www.scopus.com/inward/record.uri?eid=2-s2.0-85098669794&doi=10.1109%2fICIP40778.2020.9190802&partnerID=40&md5=43eb83516d7b49991e7ba9687b2a9f47 https://scholars.lib.ntu.edu.tw/handle/123456789/581451 |
ISSN: | 15224880 | DOI: | 10.1109/ICIP40778.2020.9190802 |
顯示於: | 資訊工程學系 |
在 IR 系統中的文件,除了特別指名其著作權條款之外,均受到著作權保護,並且保留所有的權利。