← 返回 Skills 市场

Annotation Format Converter

Name: Annotation Format Converter
Author: mingo-318

作者 Mingo_318 · GitHub ↗ · v1.0.0

cross-platform ✓ 安全检测通过

292

总下载

当前安装

版本数

在 OpenClaw 中安装

/install annotation-format-converter

功能描述

Convert annotation files between COCO, YOLO, VOC, and LabelMe formats with auto-detection and support for batch processing of folders.

使用说明 (SKILL.md)

Annotation Format Converter

Convert annotation formats between COCO, YOLO, VOC, and LabelMe. Use when user needs to convert annotation files between different formats for computer vision tasks.

Features

COCO → YOLO: Convert COCO JSON to YOLO txt format
YOLO → COCO: Convert YOLO txt to COCO JSON
VOC → COCO: Convert Pascal VOC XML to COCO JSON
LabelMe → COCO: Convert LabelMe JSON to COCO JSON
Auto-detect: Automatically detect input format
Batch Convert: Convert entire folders

Usage

# Convert COCO JSON to YOLO
python scripts/converter.py coco2yolo input.json output_dir/

# Convert YOLO txt to COCO
python scripts/converter.py yolo2coco input_dir/ output.json

# Convert VOC XML to COCO
python scripts/converter.py voc2coco input_dir/ output.json

# Auto-detect and convert
python scripts/converter.py convert input.json output.json --from coco --to yolo

# List supported formats
python scripts/converter.py formats

Supported Formats

Format	Extension	Description
COCO	.json	COCO JSON annotation
YOLO	.txt	YOLO darknet format
VOC	.xml	Pascal VOC XML
LabelMe	.json	LabelMe JSON

Examples

COCO to YOLO

$ python scripts/converter.py coco2yolo annotations.json yolo_labels/

Converting COCO to YOLO...
✓ Converted 150 annotations to yolo_labels/

YOLO to COCO

$ python scripts/converter.py yolo2coco labels/ output.json --image-dir images/

Converting YOLO to COCO...
✓ Converted 150 annotations to output.json

Installation

pip install pillow tqdm

Requirements

Python 3.8+
Pillow (for image dimensions)
tqdm (for progress bar)

安全使用建议

This skill appears to do what it says: convert annotation formats locally. Before installing/running: (1) review and run the script in a controlled directory (it will read image/annotation files you point it to and write outputs, potentially overwriting existing files); (2) install Pillow if you need image-dimension detection; tqdm is optional because the script provides a simple fallback; (3) be aware default category handling is simplistic (defaults to a single generic 'object' category and 1-index/0-index ID adjustments), so verify category mapping for your dataset; (4) if you plan to run it on sensitive directories, inspect the code yourself (it contains only local file I/O and no network/execution of remote code).

功能分析

Type: OpenClaw Skill Name: annotation-format-converter Version: 1.0.0 The skill bundle is a legitimate utility for converting computer vision annotation formats (COCO, YOLO, VOC, and LabelMe). The core logic in `scripts/converter.py` uses standard Python libraries and the Pillow library for image processing, with no evidence of network activity, data exfiltration, or malicious execution. All instructions in `SKILL.md` and the provided test files are consistent with the stated purpose of the tool.

能力评估

✓ Purpose & Capability

Name/description (convert between COCO, YOLO, VOC, LabelMe) align with the included script and test files. The script implements conversions and batch processing as advertised.

✓ Instruction Scope

SKILL.md instructs running the local Python script on local files/dirs. The script only reads annotation files and optional images to get dimensions, and writes converted annotation outputs — it does not access network endpoints or unrelated system configuration.

ℹ Install Mechanism

No install spec is provided (instruction-only), which is lowest-risk. SKILL.md recommends 'pip install pillow tqdm' — pillow is required by the script; tqdm is suggested but the script defines an internal tqdm fallback, so installing tqdm is optional.

✓ Credentials

The skill does not request environment variables, credentials, or config paths. All file and image accesses are local and proportional to an annotation-conversion tool.

✓ Persistence & Privilege

always is false and the skill does not request persistent system privileges or modify other skills or system settings.

如何使用

确保已安装 OpenClaw（本地或 Docker 部署）
在对话框中输入安装命令：/install annotation-format-converter
安装完成后，直接呼叫该 Skill 的名称或使用 /annotation-format-converter 触发
根据 Skill 的参数说明提供必要输入，即可获得结构化输出

版本历史

v1.0.0

Initial release of Annotation Format Converter. - Supports conversion between COCO, YOLO, VOC, and LabelMe annotation formats. - Enables auto-detection of annotation format. - Provides batch conversion for folders. - Includes CLI usage examples for all major conversions. - Lists supported formats and requirements.

元数据

Slug annotation-format-converter

版本 1.0.0

许可证 —

累计安装 0

当前安装数 0

历史版本数 1

常见问题