The Hong Kong Polytechnic University Contact‐free 3D/2D Hand Images Database (Version 2.0)

 

The hand biometric research has attracted a lot of attention from the researchers and generated tons of papers and technical reports. In order to promote the development of more user-friendly and reliable hand identification, The Hong Kong Polytechnic University has been collecting contact-free 3D hand scans since July 2008. The objective is to establish large scale hand image database acquired using completely contact-free imaging setup and make it freely available for researchers. Currently, there is no database of contact-free hand scans in the public domain. The only publicly available database of 3D hands contains 3D and 2D images of the dorsal surface of the hand. In contrast, this database includes textured  3D scans of the inner surface (palm-side) of the hand with significant pose variations, which can be utilized to further the research efforts in the area of contactless hand geometry as well as palmprint identification.

 

Brief Description

The 3D hand images in the database were acquired using a commercially available 3D digitizer Minolta VIVID 910. The database acquisition was performed in our university. The volunteers were the student, staff of our university and also included visitors from elsewhere. This database currently contains 1140 right hand images (3D and the corresponding 2D) acquired from 114 subjects. In order to introduce considerable pose variations in the database, subjects were instructed to present their hand in five different poses (refer to following figure). The five contactless hand images for the every user were acquired with the following suggested poses; frontal pose when the hand is held approximately parallel to the image plane of the scanner, hand is rotated in the clockwise direction about x-axix, hand is rotated in the counter-clockwise direction about x-axix, hand is rotated in the clockwise direction about y-axix, and finally when the hand is rotated in the counter-clockwise direction about y-axix. The  volunteers were given the freedom to pose at any angle as long as the hand is inside the imaging volume of the scanner and there is no significant overlap of fingers in the acquired images that would make it impossible to locate and separate fingers before pose correction. This is done in order to perform experiments and evaluate the performance prior to pose normalization. The acquired images have been sequentially numbered for every user with an integer identification/number. The resolution of these images (3D as well as 2D) is 640 x 480 pixels.

 

Sample Images

The entire database is being made available as it was acquired. The 2D image samples from two subjects in this database are reproduced in the following.

 

2D Images

 

 

   

Download and Copyright

The database is being made available freely for the researchers. Interested researchers should follow following steps to acquire the 'The Hong Kong Polytechnic University Contact‐free 3D/2D Hand Images Database (Version 2.0)'

   1. Visit the online application link and provide all the details in this application form;

   2. The request confirmation will be sent by email and the applicants can download the database using instructions in the email.

 

All the rights of the 'The Hong Kong Polytechnic University Contact‐free 3D/2D Hand Images Database (Version 2.0)' are reserved and commercial use/distribution of this database is strictly prohibited. All the technical reports and papers that report experimental results from this database should provide due acknowledgement/reference.

Reference

[1] Vivek Kanhangad, Ajay Kumar, and David Zhang, "Contactless and pose invariant biometric identification using hand surface , IEEE Trans. Image Processing, ” vol. 6, no. 3, pp. 1014, 1027, Sep. 2011.  

Back to Databases Page