z-x-yang/Segment-and-Track-Anything

An open-source project dedicated to tracking and segmenting any objects in videos, either automatically or interactively. The primary algorithms utilized...

View on GitHub ↗Jump to charts ↓Open shareable report →

Data as of . Signed-in members get hourly updates — create a free account.

Summary Information

Updated 2 hours ago
Added to GitGenius on September 22nd, 2026
Created on April 16th, 2023
Open Issues & Pull Requests: 42 (+0)
GitHub issues: Enabled
Number of forks: 355
Total Stargazers: 3,138 (+0)
Total Subscribers: 39 (+0)

Repository Insights (GitGenius)

Median issue/PR response: 37.2 days
Mean response time: 231.4 days
90th percentile: 762.1 days
Tracked items: 44

Charts & Analytics

Fetching additional details & charts...

Issue Activity (beta)

Open issues: 13
New in 7 days: 0
Closed in 7 days: 0
Avg open age: 832 days
Stale 30+ days: 12
Stale 90+ days: 1

Recent activity

Opened in 7 days: 0
Closed in 7 days: 0
Comments in 7 days: 0
Events in 7 days: 0

Top labels

No label distribution available yet.

Most active issues this week

Sign in to see which issues are moving.
Sign in

Detailed Description

Segment-and-Track-Anything is a video object segmentation tool that automatically or interactively tracks and segments arbitrary objects across video frames.

The tool addresses the challenge of maintaining consistent object identity and boundaries throughout video sequences. It combines two complementary approaches: the Segment Anything Model handles segmentation on key frames, identifying object masks with minimal prompting, while Associating Objects with Transformers manages temporal propagation, efficiently tracking those objects across subsequent frames. This two-stage design allows the system to handle both automatic processing and interactive refinement, where users can guide segmentation on specific frames and have the tool propagate results forward.

The project suits developers working on video analysis tasks where object-level understanding is required, such as video editing, surveillance analysis, or dataset annotation. It is particularly valuable when you need to segment multiple arbitrary objects without training on domain-specific data. The interactive mode makes it accessible for cases where fully automatic segmentation is insufficient, allowing manual correction on key frames while the tracking component handles the temporal consistency burden.

Development activity shows consistent engagement with the codebase. The project receives regular updates and refinements to its core algorithms. There is active responsiveness to issues raised by users, indicating ongoing maintenance. The repository demonstrates a pattern of iterative improvements rather than sporadic activity, suggesting the maintainers are invested in addressing reported problems and enhancing functionality.