Title :
1.15mW mixed-mode neuro-fuzzy accelerator for keypoint localization in image processing
Author :
Injoon Hong ; Jinwook Oh ; Hoi-Jun Yoo
Author_Institution :
Dept. of Electr. Eng., Korea Adv. Inst. of Sci. & Technol. (KAIST), Daejeon, South Korea
Abstract :
A mixed-mode neuro-fuzzy accelerator is proposed for keypoint localization of image features of Scale Invariant Feature Transform (SIFT) algorithm. To reduce processing time of keypoint localization with low power consumption, analog Adaptive Neuro-Fuzzy Inference System (ANFIS) and digital controller are implemented together. It is implemented in 0.13μm CMOS process and achieves 1.15mW power consumption. Compared to the conventional digital standalone system, 0.733mm2 neuro-fuzzy accelerator achieves 43% processing time reduction and also results in 19.4% time reduction of image feature extraction process.
Keywords :
feature extraction; fuzzy neural nets; transforms; CMOS process; analog adaptive neuro-fuzzy inference system; digital controller; digital standalone system; image feature extraction process; image processing; keypoint localization; low power consumption; mixed-mode neuro-fuzzy accelerator; power 1.15 mW; scale invariant feature transform algorithm; Noise measurement; Process control;
Conference_Titel :
Circuits and Systems (MWSCAS), 2011 IEEE 54th International Midwest Symposium on
Conference_Location :
Seoul
Print_ISBN :
978-1-61284-856-3
Electronic_ISBN :
1548-3746
DOI :
10.1109/MWSCAS.2011.6026495
