Title :
The brain mimicking Visual Attention Engine: An 80×60 digital Cellular Neural Network for rapid global feature extraction
Author :
Lee, Seungjin ; Kim, Kwanho ; Kim, Minsu ; Kim, Joo-Young ; Yoo, Hoi-Jun
Author_Institution :
Sch. of Electr. Eng. & Comput. Sci., KAIST, Daejeon
Abstract :
The visual attention engine (VAE), an 80 times 60 digital cellular neural network, rapidly extracts global features used as attentional cues to streamline detailed object recognition. A peak performance of 24 GOPS is achieved by 120 processing elements (PE) shared by the cells. 2D shift register based data transactions enable 93% PE utilization. Integrated within an object recognition SoC, the 4.5 mm2 VAE running at 200 MHz improves object recognition frame rate by 83% while consuming just 84 mW.
Keywords :
feature extraction; neural nets; object recognition; shift registers; system-on-chip; 2D shift register; GOPS; SoC; brain mimicking visual attention engine; digital cellular neural network; frequency 200 MHz; object recognition; power 84 mW; processing elements; rapid global feature extraction; Aerodynamics; Cellular neural networks; Data mining; Engines; Feature extraction; Filters; Object recognition; Routing; Shift registers; Wire;
Conference_Titel :
VLSI Circuits, 2008 IEEE Symposium on
Conference_Location :
Honolulu, HI
Print_ISBN :
978-1-4244-1804-6
Electronic_ISBN :
978-1-4244-1805-3
DOI :
10.1109/VLSIC.2008.4585938
