Title :
A 54GOPS 51.8mW analog-digital mixed mode Neural Perception Engine for fast object detection
Author :
Kim, Minsu ; Kim, Joo-Young ; Lee, Seungjin ; Oh, Jinwook ; Yoo, Hoi-Jun
Author_Institution :
Dept. of Electron. Eng. & Comput. Sci., Korea Adv. Inst. of Sci. & Technol. (KAIST), Daejeon, South Korea
Abstract :
A mixed mode Neural Perception Engine (NPE) is proposed as the pre-processing accelerator of multi-object recognition processor to reduce the computational complexity and increase its efficiency. It consists of Motion Estimator (ME), Visual Attention Engine (VAE) and Object Detection Engine (ODE). The fabricated chip achieves 54 GOPS 51.8 mW NPE. By implementing a fast and robust neuro-fuzzy algorithm in analog-digital mixed circuits, the area and power of the ODE is reduced by 59% and 44%, respectively, compared to those of all digital implementation. The NPE can increase the frame rate by 2.09x and reduce power consumption by 38% of the multi-object recognition processor.
Keywords :
fuzzy neural nets; image recognition; microprocessor chips; mixed analogue-digital integrated circuits; object detection; analog-digital mixed circuits; analog-digital mixed mode neural perception engine; computational complexity; fast object detection; motion estimator; multiobject recognition processor; neurofuzzy algorithm; object detection engine; preprocessing accelerator; visual attention engine; Analog circuits; Analog-digital conversion; Energy consumption; Engines; Heuristic algorithms; Laboratories; Motion estimation; Object detection; Object recognition; Signal processing;
Conference_Titel :
Custom Integrated Circuits Conference, 2009. CICC '09. IEEE
Conference_Location :
San Jose, CA
Print_ISBN :
978-1-4244-4071-9
Electronic_ISBN :
978-1-4244-4073-3
DOI :
10.1109/CICC.2009.5280749
