skills.rest
SharpAI
SharpAIOFFICIAL
Explore all skills by this author →
GitHub

annotation-data

Manages annotation datasets with CRUD operations, label tracking, and COCO-format export.

3.0k|478|Updated Mar 5, 2019
Search Tags:#annotation#JSONL#dataset management#bounding boxes#Machine learning#Labeling#coco export
One-click install
npx skills add https://github.com/SharpAI/DeepCamera --skill annotation-data
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: annotation-data
Source: https://github.com/SharpAI/DeepCamera/tree/main/skills/annotation/dataset-management
Command: npx skills add https://github.com/SharpAI/DeepCamera --skill annotation-data

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) components.

What problem does it solve?

Managing image annotation datasets for machine learning requires consistent storage, label tracking, and export to standard formats like COCO, which is tedious to handle manually across many frames and datasets.

Core Features & Use Cases

  • Dataset CRUD: Create, read, update, and delete annotation datasets stored as JSONL files with metadata.
  • Annotation & Label Management: Save per-frame annotations, list labels, and compute statistics such as per-label counts.
  • COCO Export: Convert datasets into standard COCO JSON format with categories, images, and bounding-box annotations for training pipelines.
  • Use Case: While labeling camera frames in an annotation studio, save bounding-box annotations per frame, track label distribution statistics, then export the finished dataset as COCO JSON for model training.

Quick Start

Ask the assistant to list my annotation datasets and export the dataset named my_dataset to COCO format.

Frequently Asked Questions about annotation-data

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I export an annotation dataset to COCO format?▼

Send the export_coco command with the dataset name over the JSONL protocol. The skill builds COCO categories from dataset labels, converts each frame record into images and annotations entries, and writes a coco_export.json file inside the dataset directory.

How do I save bounding box annotations for image frames?▼

Use the save_annotation command with the dataset name, frame_id, and a list of annotation objects containing labels and bounding boxes. Each record is appended to the dataset's annotations.jsonl file with a timestamp.

What dependencies does this annotation management tool require?▼

It requires only Python 3.9 or newer with the standard library. The requirements.txt explicitly states no external packages are needed, and the deploy script creates a plain virtual environment without installing anything.

Where are annotation datasets stored on disk?▼

Datasets are stored under a configurable datasets_dir parameter, defaulting to ~/.aegis/datasets when not specified. Each dataset is a directory containing meta.json for labels and metadata plus annotations.jsonl for frame records.

What are the limitations of JSONL-based annotation storage?▼

Annotations are append-only per frame, so updating or removing individual annotations requires rewriting files manually. There is no concurrent write locking, so simultaneous writers to the same dataset may corrupt the JSONL file.

Related Skills

Explore More Data & Analytics Skills→
facebookfacebook
506

camera-streaming

Stream live video and capture JPEG photos from Meta AI glasses in iOS apps.

Official
Intermediate
alirezarezvanialirezarezvani
25.3k

senior-computer-vision

Builds object detection, segmentation, and deployment pipelines for computer vision models.

Community
Advanced
SharpAISharpAI
3.0k

dataset-annotation

Annotate datasets with bounding boxes, SAM2 segmentation, and DINOv3 visual grounding.

Official
Intermediate
cwinkelmanncwinkelmann

active-learning-wildlife

Operationalize active learning workflows for wildlife detection with HILDA.

Community
Advanced