Separable Convolution Filters – Separable convolution filters for FPGA

Software module for high-level synthesis (HLS) of separable convolution filters on FPGA platform for Computer Vision, developed with Vivado HLS. Project for my Bachelor's degree thesis in "Electronic Calculators" at University of Bologna, supervised by professor Stefano Mattoccia.

The project constituted a step for the future development of a convolutional neural network (CNN) on the FPGA, exploiting the computational optimization of separable convolution filters and fixed-point arithmetic.

(A.Y. 2017-2018)

Documentation (italian language only)

For a brief summary of the contents of this work, read this document.

For a more detailed and in-depth description of the contents of this work, read this document.