My students and I are creating a new (freely available) database for person identification/ authentication using hand geometry and palm features, we are intending to target 200 persons, male and female, and different age groups (young, middle, and old) .
We decided to record 3 sessions for each person, 5 hand images for each person for each session, with random rotation of the hand (around the center), open the fingers randomly.
Could you please advice in this regards?
Is it enough contribution to write a paper describing the database without doing some experiments?
Could you please refer us to similar databases? because all we found available is just 10 -20 persons.
Any comment is highly appreciated.
Thank you Carlos
I was thinking in two master thesis for my students, those references are wonderful.
also we are thinking in recognizing terrorists who show only their hands
Dear @Ahmad, It is not exact my field, but I am attaching a chapter on this issue Design Aspects of Secure Biometric Systems and Biometrics in the Encrypted Domain! Hope it will be helpful!
It depends on your idea about the quality of your publications. I think if you include experiments it will have more value and the publication can be published on a better journal.
Some hand image datasets have 800 users and with different ranges, which is the different the innovation of this dataset?? If you include experiments, the publication will have that added value.
Dear Carlos
The different innovation of those data sets is that the images are taken by mobile phone camera and a webcam, we need to investigate to what extent do such low quality images capable of identifying persons.
Dear Ahmad,
It is a good idea, but It has been researched. In particular, you can find the following public dataset, The Replay-Attack Database — DDP - Idiap Research Institute (https://www.idiap.ch/dataset/replayattack) or Databases - BioSecure. And too, there are publications about it, and too, intra-modalities fusions (Intra-modal biometric system using hand-geometry and palmprint texture, JJ Fuertes, CM Travieso, MA Ferrer, JB Alonso Security Technology (ICCST), 2010 IEEE International Carnahan Conference on, pp. 318-322)
It is a bit complex to find a field on biometrics, where a researcher can make a new research. I think a good idea, it the use of multiple sensors of different resolutions and ranges (on the same time) and to fuse those informations, details and holistic information, and so, your idea of dataset can be the first step.
Kind regards,
Carlos
Thank you Carlos
I was thinking in two master thesis for my students, those references are wonderful.
also we are thinking in recognizing terrorists who show only their hands
Thank you, Ahmad.
It is a great idea, firstly, to work with two Master students to build the dataset and finally, the application, its use for identifying terrorists.
Congratulations for your good planing!!
- Badges
- Science topic
- Similar topics
- Engineering
- Electrical Engineering
- Signal Processing
- Image Processing