-
512KiB RAM Is Enough! Live Camera Face Recognition DNN on MCU
AbstractSmall factor and ultra-low power devices are becoming more and more smart and capable even for deep learning network inference. And as the devices are "small", the challenge is becoming tougher. This paper covers full development and deployment pipeline of Face Recognition with a live camera - from model training and quantization to porting to RISC-V MCU with 512 kilobytes of internal RAM. Authors provide GreenWaves GAP8 SoC overview and the approaches for DNN model optimization and inference in the extreme environment. As the project outcome, authors were able to run Face Detection and Recognition with live QVGA camera and display preview on a battery-powered board.
Related Material
[pdf][bibtex]@InProceedings{Zemlyanikin_2019_ICCV,
author = {Zemlyanikin, Maxim and Smorkalov, Alexander and Khanova, Tatiana and Petrovicheva, Anna and Serebryakov, Grigory},
title = {512KiB RAM Is Enough! Live Camera Face Recognition DNN on MCU},
booktitle = {Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV) Workshops},
month = {Oct},
year = {2019}
}
These ICCV 2019 papers are the Open Access versions, provided by the Computer Vision Foundation.
Except for the watermark, they are identical to the accepted versions; the final published version of the proceedings is available on IEEE Xplore.
Except for the watermark, they are identical to the accepted versions; the final published version of the proceedings is available on IEEE Xplore.
This material is presented to ensure timely dissemination of scholarly and technical work.
Copyright and all rights therein are retained by authors or by other copyright holders.
All persons copying this information are expected to adhere to the terms and constraints invoked by each author's copyright.