Abstract:Unsupervised learning has been popular in various computer vision tasks, including visual object tracking. However, prior unsupervised tracking approaches rely heavily on spatial supervision from template-search pairs and are still unable to track objects with strong variation over a long time span. As unlimited self-supervision signals can be obtained by tracking a video along a cycle in time, we investigate evolving a Siamese tracker by tracking videos forward-backward. We present a novel unsupervised tracking framework, in which we can learn temporal correspondence both on the classification branch and regression branch. Specifically, to propagate reliable template feature in the forward propagation process so that the tracker can be trained in the cycle, we first propose a consistency propagation transformation. We then identify an ill-posed penalty problem in conventional cycle training in backward propagation process. Thus, a differentiable region mask is proposed to select features as well as to implicitly penalize tracking errors on intermediate frames. Moreover, since noisy labels may degrade training, we propose a mask-guided loss reweighting strategy to assign dynamic weights based on the quality of pseudo labels. In extensive experiments, our tracker outperforms preceding unsupervised methods by a substantial margin, performing on par with supervised methods on large-scale datasets such as TrackingNet and LaSOT. Code is available at <a class="link-external link-https" href="https://github.com/FlorinShum/ULAST" rel="external noopener nofollow">this https URL</a>.

Contrastive Cycle Consistency Learning for Unsupervised Visual Tracking.

Exploiting Temporal Coherence for Self-Supervised Visual Tracking by Using Vision Transformer

Consistency-based Self-Supervised Visual Tracking by Using Query-Communication Transformer.

Continuity-Discrimination Convolutional Neural Network for Visual Object Tracking

Tracking without Label: Unsupervised Multiple Object Tracking via Contrastive Similarity Learning

Unsupervised cycle-consistent adversarial attacks for visual object tracking

Unifying Visual and Vision-Language Tracking via Contrastive Learning

Contrastive Learning of Image Representations with Cross-Video Cycle-Consistency

STC: Spatio-Temporal Contrastive Learning for Video Instance Segmentation.

Semi-TCL: Semi-Supervised Track Contrastive Representation Learning

Unsupervised Learning of Accurate Siamese Tracking

Learning By Analogy: Reliable Supervision From Transformations For Unsupervised Optical Flow Estimation

Joint Spatio-Temporal Similarity and Discrimination Learning for Visual Tracking

Unsupervised CD in satellite image time series by contrastive learning and feature tracking

Unsupervised Deep Tracking

Breaking Shortcut: Exploring Fully Convolutional Cycle-Consistency for Video Correspondence Learning

Contrastive Transformation for Self-supervised Correspondence Learning

Learning multi-view visual correspondences with self-supervision

Self-Supervised Multi-Object Tracking with Path Consistency

Self-Supervised Any-Point Tracking by Contrastive Random Walks