Biometrics Research Group at The Institute of Electronics and Computer Science

Image sensor module developed (2nd version)

May 20th, 2011 | Posted by rihards in Research - (Comments Off on Image sensor module developed (2nd version))

Second version of camera module for palm vein and palm print image acquisition is developed. It consists of image sensor sensitive in visible and infrared light spectrum, high speed interconnection with FPGA development kit DE2, high speed USB interface for debugging and data transfer to PC. This design also incorporates the visible/infrared filter exchange circuitry for fast and easy light spectrum switching.

Complex 2D Matched Filter algorithm

January 14th, 2011 | Posted by rihards in Research - (Comments Off on Complex 2D Matched Filter algorithm)

Complex 2D Matched filter (CMF) based on the Matched filter approach was developed for palm vein extraction from noisy images. This filtering approach enables to carry out computations more efficiently, because of the one filtering mask. Filtering mask consists of  real and imaginary part, therefore it is called complex. The result of such filtering is a matrix of vectors, where in each input image pixel the vector is constructed. Vectors represent the similarity between image and filtering mask in certain region. This vector information is significant for further processing tasks.

Example of the CMF masks 



First image sensor module based on the KODAK KAC image sensor is developed to capture palm images in visible and infrared light spectrum. Specially designed for easy connection to an ALTERA FPGA development kits with 40 pin connector. In the image you can see the white and IR LEDs that are used to capture palm print and palm blood vessel images. Also a Tamron lens that offers good light sensitivity with f/1.2 aperture.