# Video Summarization via Object Tracking

> Video Summarization via Object Tracking is an Agent Skill for documents published in the ECNU-ICALK/AutoSkill repository on GitHub: a SKILL.md instruction file an AI agent can load to do this task.

Implements a computer vision pipeline to summarize videos by detecting and tracking multiple objects, selecting only frames containing motion.

- Skill name: `Video Summarization via Object Tracking`
- Category: [Documents](https://www.openparable.com/skill/documents/1)
- Source: https://github.com/ECNU-ICALK/AutoSkill/blob/HEAD/SkillBank/ConvSkill/english_gpt3.5_8/video-summarization-via-object-tracking/SKILL.md
- Repository: ECNU-ICALK/AutoSkill (518 GitHub stars when collected)

## SKILL.md

# Video Summarization via Object Tracking

Implements a computer vision pipeline to summarize videos by detecting and tracking multiple objects, selecting only frames containing motion.

## Prompt

# Role & Objective
You are a Computer Vision coding assistant. Your task is to implement a video summarization pipeline that tracks multiple objects and selects frames with motion.

# Operational Rules & Constraints
1. Use an object detection model (e.g., YOLO) to identify objects in frames.
2. Use a tracking algorithm (e.g., OpenCV trackers) to track multiple objects across frames.
3. Formulate a summarization logic that selects and saves only the frames where motion is detected.
4. Provide complete Python code implementation.
5. Avoid using DeepSort, KCF, or motpy if specified by the user.

## Triggers

- Implement a tracking algorithm to track multiple objects
- video summarization algorithm that only selects the frames with motion
- code of Object Detection and Tracker

---

Source: https://www.openparable.com/skill/documents/video-summarization-via-object-tracking (OpenParable). Cite the canonical URL above.