2022, September

AVisT: A Benchmark for Visual Object Tracking in Adverse Visibility was accepted for publication at BMVC 2022.

2022, July

Robust Visual Tracking by Segmentation was accepted for publication at ECCV 2022.

2022, March

Transforming Model Prediction for Tracking was accepted for publication at CVPR 2022.

2021, September

Code for our tracker KeepTrack is availabe on github.

2021, July

Selected Publications

AVisT is a new video object tracking benchmark containing challenging videos in adverse visibility, such as severe weather conditions, camouflage and imaging effects.

BMVC, 2022

RTS is a segmentation-centric tracker that uses segmentation masks as internal target representation in order to produce accurate segmentation masks of the target instead of bounding boxes in each frame.

ECCV, 2022

ToMP is a novel DCF based tracker that uses a Transformer based model predictor for both target localization and bounding box estimation.

CVPR, 2022

Novel tracking method that tracks both the target and distractor object using a target candidate association network.

ICCV, 2021

Training strategy to reduced the gap between a network trained on full or on weak annotations.

CVIU, 2021

AFDA is a semi-supervised learning method that aligns feature distributions between labeled and unlabeled to boost the classification accuracy.

CVIU, 2021

Novel network compression techique that unifies filter pruning and low-rank decomposition employing group sparsity regularization.

CVPR, 2020

Efficient Video Segmentation (EVS) pipeline that combines a fast CPU optical flow method with two CNNs to predict dense semantic labels.

WACV, 2020

ASAL is a new active learning strategy that synthesizes uncertain samples instead of performing an exhaustive search in each active learning cycle.

WACV, 2020

Convolutional network accelerator which is able to handle GPU-CNN sizes within the power envelope of embedded systems.