Dataset assetOpen Source CommunityGesture RecognitionImage Data

cj-mills/hagrid-sample-120k-384p

The HaGRID sample dataset contains 127,331 images at 384p resolution, sampled from the original HaGRID dataset (552,992 images at 1080p, total size 716 GB). It was created to enable usage on free tiers of Google Colab and Kaggle Notebooks. The dataset is primarily intended for object‑detection tasks and includes annotations for multiple hand‑gesture categories such as call, no_gesture, like, etc. Annotation data comprise bounding‑box coordinates and class labels.

Source

hugging_face

Created

Nov 28, 2025

Updated

Jul 2, 2023

Signals

221 views

Availability

Linked source ready

Overview

Dataset description and usage context

Dataset Overview

Name: HaGRID Sample 120k 384p
License: CC‑BY‑SA‑4.0
Task Category: Object Detection
Language: English
Size Category: 100 K < samples < 1 M

Dataset Content

Image Count: 127,331
Image Resolution: 384p
Original Dataset Size: 716 GB, 552,992 images at 1080p

Object Categories

[call,
 no_gesture,
 dislike,
 fist,
 four,
 like,
 mute,
 ok,
 one,
 palm,
 peace,
 peace_inverted,
 rock,
 stop,
 stop_inverted,
 three,
 three2,
 two_up,
 two_up_inverted]

Annotation Information

Format: bboxes: [top‑left‑X, top‑left‑Y, width, height]
Coordinate Adjustment: Multiply top‑left‑X and width by the image width; multiply top‑left‑Y and height by the image height.

Example Annotation

Field	Value
bboxes	`[[0.23925175, 0.28595301, 0.25055143, 0.20777627]]`
labels	`[call]`
leading_hand	`right`
leading_conf	`1`
user_id	`5a389ffe1bed6660a59f4586c7d8fe2770785e5bf79b09334aa951f6f119c024`

Need downstream help?

Pair the dataset with AI analysis and content workflows.

Once the source passes your review, move straight into summarization, transformation, report drafting, or presentation generation with the JuheAI toolchain.

Explore AI studio