Sign Language Image Dataset Creation

Title: Create a Custom Sign Language Image Dataset (Alphabets, Numbers, and 20 Word Gestures)
Project Overview
I am developing a computer vision model that recognizes complete sign language gestures, including alphabet signs, number signs, and common word gestures.
To train the model, I need a custom image dataset of sign language gestures captured as still images. The dataset should use one consistent sign language system (for example, American Sign Language – ASL). Please specify which sign language you will use.
Only images are required. No video is needed.
Dataset Requirements
The dataset should include:
26 alphabet signs (A–Z)
10 number signs (0–9)
20 common word gestures
Total gesture classes:
56 gesture classes
Example word gestures may include:
hello
thank_you
please
sorry
help
stop
yes
no
good
bad
call
eat
drink
come
go
wait
love
think
money
finish
Image Count
Minimum requirement:
100 images per gesture class
Estimated dataset size:
56 classes × 50 images ≈ 2800 images
Images must be:
720p resolution or higher
clear and well-lit
focused primarily on the hand gesture