A Near-Sensor Processing Accelerator for Approximate Local Binary Pattern Networks
Document Type
Article
Publication Date
1-1-2024
Abstract
In this work, a high-speed and energy-efficient comparator-based Near-Sensor Local Binary Pattern accelerator architecture (NS-LBP) is proposed to execute a novel local binary pattern deep neural network. First, inspired by recent LBP networks, we design an approximate, hardware-oriented, and multiply-accumulate (MAC)-free network named Ap-LBP for efficient feature extraction, further reducing the computation complexity. Then, we develop NS-LBP as a processing-in-SRAM unit and a parallel in-memory LBP algorithm to process images near the sensor in a cache, remarkably reducing the power consumption of data transmission to an off-chip processor. Our circuit-to-application co-simulation results on MNIST and SVHN datasets demonstrate minor accuracy degradation compared to baseline CNN and LBP-network models, while NS-LBP achieves 1.25 GHz and an energy-efficiency of 37.4 TOPS/W. NS-LBP reduces energy consumption by 2.2× and execution time by a factor of 4× compared to the best recent LBP-based networks.
Identifier
85162635288 (Scopus)
Publication Title
IEEE Transactions on Emerging Topics in Computing
External Full Text Location
https://doi.org/10.1109/TETC.2023.3285493
e-ISSN
21686750
First Page
73
Last Page
83
Issue
1
Volume
12
Grant
2228028
Fund Ref
National Science Foundation
Recommended Citation
Angizi, Shaahin; Morsali, Mehrdad; Tabrizchi, Sepehr; and Roohi, Arman, "A Near-Sensor Processing Accelerator for Approximate Local Binary Pattern Networks" (2024). Faculty Publications. 1164.
https://digitalcommons.njit.edu/fac_pubs/1164