HiRISE: High-Resolution Image Scaling for Edge ML via In-Sensor Compression and Selective ROI

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Reidy, Brendan, Tabrizchi, Sepehr, Mohammadi, Mohamadreza, Angizi, Shaahin, Roohi, Arman, Zand, Ramtin
Natura: Preprint
Pubblicazione: 2024
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866911981662896128
author Reidy, Brendan
Tabrizchi, Sepehr
Mohammadi, Mohamadreza
Angizi, Shaahin
Roohi, Arman
Zand, Ramtin
author_facet Reidy, Brendan
Tabrizchi, Sepehr
Mohammadi, Mohamadreza
Angizi, Shaahin
Roohi, Arman
Zand, Ramtin
contents With the rise of tiny IoT devices powered by machine learning (ML), many researchers have directed their focus toward compressing models to fit on tiny edge devices. Recent works have achieved remarkable success in compressing ML models for object detection and image classification on microcontrollers with small memory, e.g., 512kB SRAM. However, there remain many challenges prohibiting the deployment of ML systems that require high-resolution images. Due to fundamental limits in memory capacity for tiny IoT devices, it may be physically impossible to store large images without external hardware. To this end, we propose a high-resolution image scaling system for edge ML, called HiRISE, which is equipped with selective region-of-interest (ROI) capability leveraging analog in-sensor image scaling. Our methodology not only significantly reduces the peak memory requirements, but also achieves up to 17.7x reduction in data transfer and energy consumption.
format Preprint
id arxiv_https___arxiv_org_abs_2408_03956
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle HiRISE: High-Resolution Image Scaling for Edge ML via In-Sensor Compression and Selective ROI
Reidy, Brendan
Tabrizchi, Sepehr
Mohammadi, Mohamadreza
Angizi, Shaahin
Roohi, Arman
Zand, Ramtin
Computer Vision and Pattern Recognition
With the rise of tiny IoT devices powered by machine learning (ML), many researchers have directed their focus toward compressing models to fit on tiny edge devices. Recent works have achieved remarkable success in compressing ML models for object detection and image classification on microcontrollers with small memory, e.g., 512kB SRAM. However, there remain many challenges prohibiting the deployment of ML systems that require high-resolution images. Due to fundamental limits in memory capacity for tiny IoT devices, it may be physically impossible to store large images without external hardware. To this end, we propose a high-resolution image scaling system for edge ML, called HiRISE, which is equipped with selective region-of-interest (ROI) capability leveraging analog in-sensor image scaling. Our methodology not only significantly reduces the peak memory requirements, but also achieves up to 17.7x reduction in data transfer and energy consumption.
title HiRISE: High-Resolution Image Scaling for Edge ML via In-Sensor Compression and Selective ROI
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2408.03956