DWSR: an architecture optimization framework for adaptive super-resolution neural networks based on meta-heuristics

Kód výsledku v IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F61989100%3A27240%2F24%3A10256846" target="_blank" >RIV/61989100:27240/24:10256846 - isvavai.cz</a>

Výsledek na webu

<a href="https://link.springer.com/article/10.1007/s10462-023-10648-4" target="_blank" >https://link.springer.com/article/10.1007/s10462-023-10648-4</a>

DOI - Digital Object Identifier

<a href="http://dx.doi.org/10.1007/s10462-023-10648-4" target="_blank" >10.1007/s10462-023-10648-4</a>

Jazyk výsledku

angličtina

Název v původním jazyce

DWSR: an architecture optimization framework for adaptive super-resolution neural networks based on meta-heuristics

Popis výsledku v původním jazyce

Despite recent advancements in super-resolution neural network optimization, a fundamental challenge remains unresolved: as the number of parameters is reduced, the network&apos;s performance significantly deteriorates. This paper presents a novel framework called the Depthwise Separable Convolution Super-Resolution Neural Network Framework (DWSR) for optimizing super-resolution neural network architectures. The depthwise separable convolutions are introduced to reduce the number of parameters and minimize the impact on the performance of the super-resolution neural network. The proposed framework uses the RUNge Kutta optimizer (RUN) variant (MoBRUN) as the search method. MoBRUN is a multi-objective binary version of RUN, which balances multiple objectives when optimizing the neural network architecture. Experimental results on publicly available datasets indicate that the DWSR framework can reduce the number of parameters of the Residual Dense Network (RDN) model by 22.17% while suffering only a minor decrease of 0.018 in Peak Signal-to-Noise Ratio (PSNR), the framework can reduce the number of parameters of the Enhanced SRGAN (ESRGAN) model by 31.45% while losing only 0.08 PSNR. Additionally, the framework can reduce the number of parameters of the HAT model by 5.38% while losing only 0.02 PSNR.

Název v anglickém jazyce

DWSR: an architecture optimization framework for adaptive super-resolution neural networks based on meta-heuristics

Popis výsledku anglicky

Despite recent advancements in super-resolution neural network optimization, a fundamental challenge remains unresolved: as the number of parameters is reduced, the network&apos;s performance significantly deteriorates. This paper presents a novel framework called the Depthwise Separable Convolution Super-Resolution Neural Network Framework (DWSR) for optimizing super-resolution neural network architectures. The depthwise separable convolutions are introduced to reduce the number of parameters and minimize the impact on the performance of the super-resolution neural network. The proposed framework uses the RUNge Kutta optimizer (RUN) variant (MoBRUN) as the search method. MoBRUN is a multi-objective binary version of RUN, which balances multiple objectives when optimizing the neural network architecture. Experimental results on publicly available datasets indicate that the DWSR framework can reduce the number of parameters of the Residual Dense Network (RDN) model by 22.17% while suffering only a minor decrease of 0.018 in Peak Signal-to-Noise Ratio (PSNR), the framework can reduce the number of parameters of the Enhanced SRGAN (ESRGAN) model by 31.45% while losing only 0.08 PSNR. Additionally, the framework can reduce the number of parameters of the HAT model by 5.38% while losing only 0.02 PSNR.

Druh

J_imp - Článek v periodiku v databázi Web of Science

CEP obor

—

OECD FORD obor

10200 - Computer and information sciences

Projekt

—

Návaznosti

S - Specificky vyzkum na vysokych skolach

Rok uplatnění

2024

Kód důvěrnosti údajů

S - Úplné a pravdivé údaje o projektu nepodléhají ochraně podle zvláštních právních předpisů

Název periodika

Artificial Intelligence Review

ISSN

0269-2821

e-ISSN

1573-7462

Svazek periodika

57

Číslo periodika v rámci svazku

2

Stát vydavatele periodika

US - Spojené státy americké

Počet stran výsledku

21

Strana od-do

—

Kód UT WoS článku

001161054000003

EID výsledku v databázi Scopus

—