Automated deep-learning model optimization framework for microcontrollers

This paper introduces a framework for optimizing deep-learning models on microcontrollers (MCUs) that is crucial in today’s expanding embedded device market. We focus on model optimization techniques, particularly pruning and quantization, to enhance the performance of neural networks within the lim...

Full description

Saved in:
Bibliographic Details
Main Authors: Seungtae Hong, Gunju Park, Jeong-Si Kim
Format: Article
Language:English
Published: Electronics and Telecommunications Research Institute (ETRI) 2025-04-01
Series:ETRI Journal
Subjects:
Online Access:https://doi.org/10.4218/etrij.2023-0522
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849725653285339136
author Seungtae Hong
Gunju Park
Jeong-Si Kim
author_facet Seungtae Hong
Gunju Park
Jeong-Si Kim
author_sort Seungtae Hong
collection DOAJ
description This paper introduces a framework for optimizing deep-learning models on microcontrollers (MCUs) that is crucial in today’s expanding embedded device market. We focus on model optimization techniques, particularly pruning and quantization, to enhance the performance of neural networks within the lim-ited resources of MCUs. Our approach combines automatic iterative optimization and code generation, simplifying MCU model deployment without requiring extensive hardware knowledge. Based on experiments with architec-tures, such as ResNet-8 and MobileNet v2, our framework substantially reduces the model size and enhances inference speed that are crucial for MCU efficiency. Compared with TensorFlow Lite for MCUs, our optimizations for MobileNet v2 reduce static random-access memory use by 51%-57% and flash use by 17%-62%, while increasing inference speed by approximately 1.55 times. These advancements highlight the impact of our method on perfor-mance and memory efficiency, demonstrating its value in embedded artificial intelligence and broad applicability in MCU-based neural network optimization.
format Article
id doaj-art-a7309bb488bd406090fcdffd735fcfcb
institution DOAJ
issn 1225-6463
2233-7326
language English
publishDate 2025-04-01
publisher Electronics and Telecommunications Research Institute (ETRI)
record_format Article
series ETRI Journal
spelling doaj-art-a7309bb488bd406090fcdffd735fcfcb2025-08-20T03:10:25ZengElectronics and Telecommunications Research Institute (ETRI)ETRI Journal1225-64632233-73262025-04-0147217919210.4218/etrij.2023-0522Automated deep-learning model optimization framework for microcontrollersSeungtae HongGunju ParkJeong-Si KimThis paper introduces a framework for optimizing deep-learning models on microcontrollers (MCUs) that is crucial in today’s expanding embedded device market. We focus on model optimization techniques, particularly pruning and quantization, to enhance the performance of neural networks within the lim-ited resources of MCUs. Our approach combines automatic iterative optimization and code generation, simplifying MCU model deployment without requiring extensive hardware knowledge. Based on experiments with architec-tures, such as ResNet-8 and MobileNet v2, our framework substantially reduces the model size and enhances inference speed that are crucial for MCU efficiency. Compared with TensorFlow Lite for MCUs, our optimizations for MobileNet v2 reduce static random-access memory use by 51%-57% and flash use by 17%-62%, while increasing inference speed by approximately 1.55 times. These advancements highlight the impact of our method on perfor-mance and memory efficiency, demonstrating its value in embedded artificial intelligence and broad applicability in MCU-based neural network optimization. https://doi.org/10.4218/etrij.2023-0522automated frameworkdeep learningmemory efficiencymicrocontrollersmodel optimization
spellingShingle Seungtae Hong
Gunju Park
Jeong-Si Kim
Automated deep-learning model optimization framework for microcontrollers
ETRI Journal
automated framework
deep learning
memory efficiency
microcontrollers
model optimization
title Automated deep-learning model optimization framework for microcontrollers
title_full Automated deep-learning model optimization framework for microcontrollers
title_fullStr Automated deep-learning model optimization framework for microcontrollers
title_full_unstemmed Automated deep-learning model optimization framework for microcontrollers
title_short Automated deep-learning model optimization framework for microcontrollers
title_sort automated deep learning model optimization framework for microcontrollers
topic automated framework
deep learning
memory efficiency
microcontrollers
model optimization
url https://doi.org/10.4218/etrij.2023-0522
work_keys_str_mv AT seungtaehong automateddeeplearningmodeloptimizationframeworkformicrocontrollers
AT gunjupark automateddeeplearningmodeloptimizationframeworkformicrocontrollers
AT jeongsikim automateddeeplearningmodeloptimizationframeworkformicrocontrollers