We introduce a novel framework called RefineVIS for Video Instance
Segme...
Video instance segmentation aims at predicting object segmentation masks...
Improving the generalization capability of Deep Neural Networks (DNNs) i...
Unsupervised domain adaptive person re-identification (ReID) has been
ex...
Multi-camera tracking systems are gaining popularity in applications tha...
Tracking multiple objects in videos relies on modeling the spatial-tempo...
Multiple Object Tracking (MOT) has witnessed remarkable advances in rece...
Despite great recent advances in visual tracking, its further developmen...
Real-world data often follow a long-tailed distribution as the frequency...
There has been a long pursuit for precise and reproducible glomerular
qu...
Development of next-generation electronic devices for applications call ...
Generic visual tracking is difficult due to many challenge factors (e.g....
Detecting objects in aerial images is challenging for at least two reaso...
Data association-based multiple object tracking (MOT) involves multiple
...
Recent progresses in model-free single object tracking (SOT) algorithms ...
Recurrent neural networks (RNNs) have shown the ability to improve scene...
In this paper, we present LaSOT, a high-quality benchmark for Large-scal...