Finger Tapping Hand Keypoints Dataset (HKD)

Srujana Gattupalli1, Ashwin Ramesh Babu2
1Vision-Learning-Mining Laborartory, 2Heracleia Human-Computer Inteaction Lab
The Universiy of Texas at Arlington, TX, USA

 

Overview
   
We present the Hand Keypoints Dataset (HKD) with images from participants performing gesture corresponding to "finger appose" and "appose finger sucession". These are hand poses correspond to certain exercises for cognitive behavior assessment and their gestures are as follows:

Appose Finger Sucession: In this exercise, the participant is instructed to tap each finger against the thumb in order of Thumb to Index, Middle, Ring and Little finger respectively. This sequence is to be repeated as as fast as possible. The participant is told not to tap in the backwards order.

Finger Tapping: The participant is instructed to tap index finger against the thumb as fast as possible.

We provide annotations for 6 Keypoint locations on the hand namely - Wrist, Thumb, Index, Middle, Ring and Little finger tips. There are original images 640 X 480 and zoomed in cropped images 257 x 257. The cropped images are losely cropped.

Keypoint annotations are provided for original as well as cropped images.

Example annotations of cropped images:

 


Cropped Images

   
Our cropped images HKD dataset comprises of 782 images from the exercise performances from four different subjects and annotations for them. These images are of 257 X 257 and are lossely cropped right hands from the participants' captured RGB frames. We have not provided images here due to copyright issues and can be requested by emailing authors. The annotations tool to display annotations and tool is provided in section below.

HKD_Cropped_Dataset

 
Original Images
   
Our original images HKD dataset comprises of 782 images from the exercise performances from four different subjects and annotations for them. These images are of 640 X 480 and have annotations for right hands from the participants' captured RGB frames. These images and annotations can be requested by emailing the authors.


Annotations

Annotations Tool

The tool that we created to annotate our videos - extract RGB images is provided at our Github link: https://github.com/gsrujana/Hand_Anntotaion_Tool

This tool extracts frames from .avi videos only. For each frame click on centroid of hand and it would zoom into 256 x 256 region for that centroid. Now each finger tip and wrist can be annotated. Save button saves the annotations in original image coordinates and annotations translated to cropped images. Original and cropped images get saved in different folders.

 

 


Citation:
If you use our dataset for your research work, please Cite our Paper:
"Towards Deep Learning based Hand Keypoints Detection for Rapid Sequential Movements from RGB images ."

Thank You for Visiting our Webpage!