Simultaneous Learning Knowledge Distillation for Image Restoration: Efficient Model Compression for Drones

Deploying high-performance image restoration models on drones is critical for applications like autonomous navigation, surveillance, and environmental monitoring. However, the computational and memory limitations of drones pose significant challenges to utilizing complex image restoration models in...

Full description

Saved in:
Bibliographic Details
Main Author: Yongheng Zhang
Format: Article
Language:English
Published: MDPI AG 2025-03-01
Series:Drones
Subjects:
Online Access:https://www.mdpi.com/2504-446X/9/3/209
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850089896996241408
author Yongheng Zhang
author_facet Yongheng Zhang
author_sort Yongheng Zhang
collection DOAJ
description Deploying high-performance image restoration models on drones is critical for applications like autonomous navigation, surveillance, and environmental monitoring. However, the computational and memory limitations of drones pose significant challenges to utilizing complex image restoration models in real-world scenarios. To address this issue, we propose the <b>S</b>imultaneous <b>L</b>earning <b>K</b>nowledge <b>D</b>istillation (<b>SLKD</b>) framework, specifically designed to compress image restoration models for resource-constrained drones. SLKD introduces a dual-teacher, single-student architecture that integrates two complementary learning strategies: <b>D</b>egradation <b>R</b>emoval <b>L</b>earning (<b>DRL</b>) and <b>I</b>mage <b>R</b>econstruction <b>L</b>earning (<b>IRL</b>). In DRL, the student encoder learns to eliminate degradation factors by mimicking Teacher A, which processes degraded images utilizing a BRISQUE-based extractor to capture degradation-sensitive natural scene statistics. Concurrently, in IRL, the student decoder reconstructs clean images by learning from Teacher B, which processes clean images, guided by a PIQE-based extractor that emphasizes the preservation of edge and texture features essential for high-quality reconstruction. This dual-teacher approach enables the student model to learn from both degraded and clean images simultaneously, achieving robust image restoration while significantly reducing computational complexity. Experimental evaluations across five benchmark datasets and three restoration tasks—deraining, deblurring, and dehazing—demonstrate that, compared to the teacher models, the SLKD student models achieve an average reduction of 85.4% in FLOPs and 85.8% in model parameters, with only a slight average decrease of 2.6% in PSNR and 0.9% in SSIM. These results highlight the practicality of integrating SLKD-compressed models into autonomous systems, offering efficient and real-time image restoration for aerial platforms operating in challenging environments.
format Article
id doaj-art-d47aab5a207f430980f8802b5eb72d3a
institution DOAJ
issn 2504-446X
language English
publishDate 2025-03-01
publisher MDPI AG
record_format Article
series Drones
spelling doaj-art-d47aab5a207f430980f8802b5eb72d3a2025-08-20T02:42:40ZengMDPI AGDrones2504-446X2025-03-019320910.3390/drones9030209Simultaneous Learning Knowledge Distillation for Image Restoration: Efficient Model Compression for DronesYongheng Zhang0School of Computer, Beijing University of Posts and Telecommunications, No. 10 Xitucheng Road, Haidian District, Beijing 100876, ChinaDeploying high-performance image restoration models on drones is critical for applications like autonomous navigation, surveillance, and environmental monitoring. However, the computational and memory limitations of drones pose significant challenges to utilizing complex image restoration models in real-world scenarios. To address this issue, we propose the <b>S</b>imultaneous <b>L</b>earning <b>K</b>nowledge <b>D</b>istillation (<b>SLKD</b>) framework, specifically designed to compress image restoration models for resource-constrained drones. SLKD introduces a dual-teacher, single-student architecture that integrates two complementary learning strategies: <b>D</b>egradation <b>R</b>emoval <b>L</b>earning (<b>DRL</b>) and <b>I</b>mage <b>R</b>econstruction <b>L</b>earning (<b>IRL</b>). In DRL, the student encoder learns to eliminate degradation factors by mimicking Teacher A, which processes degraded images utilizing a BRISQUE-based extractor to capture degradation-sensitive natural scene statistics. Concurrently, in IRL, the student decoder reconstructs clean images by learning from Teacher B, which processes clean images, guided by a PIQE-based extractor that emphasizes the preservation of edge and texture features essential for high-quality reconstruction. This dual-teacher approach enables the student model to learn from both degraded and clean images simultaneously, achieving robust image restoration while significantly reducing computational complexity. Experimental evaluations across five benchmark datasets and three restoration tasks—deraining, deblurring, and dehazing—demonstrate that, compared to the teacher models, the SLKD student models achieve an average reduction of 85.4% in FLOPs and 85.8% in model parameters, with only a slight average decrease of 2.6% in PSNR and 0.9% in SSIM. These results highlight the practicality of integrating SLKD-compressed models into autonomous systems, offering efficient and real-time image restoration for aerial platforms operating in challenging environments.https://www.mdpi.com/2504-446X/9/3/209knowledge distillationmodel compressiondrone-view image restoration
spellingShingle Yongheng Zhang
Simultaneous Learning Knowledge Distillation for Image Restoration: Efficient Model Compression for Drones
Drones
knowledge distillation
model compression
drone-view image restoration
title Simultaneous Learning Knowledge Distillation for Image Restoration: Efficient Model Compression for Drones
title_full Simultaneous Learning Knowledge Distillation for Image Restoration: Efficient Model Compression for Drones
title_fullStr Simultaneous Learning Knowledge Distillation for Image Restoration: Efficient Model Compression for Drones
title_full_unstemmed Simultaneous Learning Knowledge Distillation for Image Restoration: Efficient Model Compression for Drones
title_short Simultaneous Learning Knowledge Distillation for Image Restoration: Efficient Model Compression for Drones
title_sort simultaneous learning knowledge distillation for image restoration efficient model compression for drones
topic knowledge distillation
model compression
drone-view image restoration
url https://www.mdpi.com/2504-446X/9/3/209
work_keys_str_mv AT yonghengzhang simultaneouslearningknowledgedistillationforimagerestorationefficientmodelcompressionfordrones