Skip to content
conftrace_
2021 CVPR CVPR 2021

Track To Detect and Segment: An Online Multi-Object Tracker

Abstract

Most online multi-object trackers perform object detection stand-alone in a neural net without any input from tracking. In this paper, we present a new online joint detection and tracking model, TraDeS (TRAck to DEtect and Segment), exploiting tracking clues to assist detection end-to-end. TraDeS infers object tracking offset by a cost volume, which is used to propagate previous object features for improving current object detection and segmentation. Effectiveness and superiority of TraDeS are shown on 4 datasets, including MOT (2D tracking), nuScenes (3D tracking), MOTS and Youtube-VIS (instance segmentation tracking). Project page: https://jialianwu.com/projects/TraDeS.html.

🌉 Interdisciplinary Bridge - Artificial Intelligence and Computer Vision
🐣 Hot Topic Early Bird - cost volume

Authors

Jialian Wu, Jiale Cao, Liangchen Song, Yu Wang, Ming Yang, Junsong Yuan

Topics

Computer Vision > Analysis > Object Detection Computer Vision > Analysis > Object Tracking Computer Vision > Analysis > Semantic Segmentation Computer Vision > Processing > Video Understanding Computer Vision > Analysis > Object Segmentation Deep Learning > Learning Types > Multi-Task Learning

Keywords

object detection instance segmentation video tracking multi-object tracking visual tracking feature propagation cost volume multi-object tracker tracking offset
Download PDF

More from CVPR 2021

All CVPR 2021 papers →
Domain-Robust VQA With Diverse Datasets and Methods but No Target Labels
(AF)2-S3Net: Attentive Feature Fusion With Adaptive Feature Selection for Sparse Semantic Segmentation Network
Towards Real-World Blind Face Restoration With Generative Facial Prior
Look Before You Speak: Visually Contextualized Utterances
DivCo: Diverse Conditional Image Synthesis via Contrastive Generative Adversarial Network
Effective Sparsification of Neural Networks With Global Sparsity Constraint