Recognition FPGA system for detection of anomalies in mammograms

The recognizing of cancer in mammograms implemented on the FPGA platform is pretty much similar to PC result evaluation. With less than 0.6% error, this program can be used in other image processing applications that require the neural networks employment with good results in characteristic recognizing of the images. The FPGA is used as a platform to implement the rank filter and the neural network hardware employing a scheme off-line training (by software). Last facilitates the implementation due to the ability of reprogramming the FPGA. In this case, the program could be implemented successfully in the Stratix EP1S80B956C6 FPGA using less than 15% of its resources.