On an Improved FPGA Implementation of CNN-Based Gabor-Type Filters

• Published in: IEEE transactions on circuits and systems. II, Express briefs Vol. 59; no. 11; pp. 815 - 819
• Main Authors: Cesur, E., Yildiz, N., Tavsanoglu, V.
• Format: Journal Article
• Language: English
• Published: New York IEEE 01.11.2012; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Architecture; Bandwidth; Cellular neural networks (CNNs); Circuits; Computer architecture; Convergence; Equations; Field programmable gate arrays; field-programmable gate arrays (FPGAs); Finite impulse response filter; Gabor filters; Neural networks; Pixels; Real time; real-time systems; reconfigurable architectures; Streams
• ISSN: 1549-7747; 1558-3791
• DOI: 10.1109/TCSII.2012.2218471

In this brief, the details of the architecture of a previously introduced improved field-programmable gate array implementation of the cellular neural network (CNN)-based 2-D Gabor-type filter are given, and the implementation results are discussed. The proposed architecture is suitable for real-time applications with high pixel rates. The prototype is capable of processing video streams up to a pixel rate of 373.2 megapixels per second (MP/s), including full-high-definition (HD) 1080p@60 (1080 × 1920 resolution, 60-Hz frame rate, and 124.4-MP/s visible pixel rate). This brief also contains convergence rate analysis results, along with some discussions on FIR and CNN-based implementation methods.