분류
2025년 2월
작성일
2024.10.10
수정일
2025.02.05
작성자
등 제강
조회수
82

A Low-cost Deep Learning Model for Real-time Low Light Image Enhancement and Defogging

Artificial Intelligence (AI) has redefined long-standing challenges in image processing, particularly in defogging and low-light image enhancement. The inherent flexibility and adaptability of AI technologies have significantly advanced enhancement outcomes. However, deploying state-of-the-art image enhancement models on low-power devices while maintaining real-time processing capabilities remains a critical challenge. This paper introduces a novel AI model architecture with approximately 4K parameters, developed using deep learning techniques and grounded in well-established principles such as the Atmospheric Scattering Model, Dark Channel Prior, and Local Maximum Color Value Prior. The proposed model comprises three key components: a transmission map estimation module, a color correction module,and a denoising module. Inspired by advanced concepts such as Residual Networks and Feature Pyramid Networks, the architecture achieves an optimal balance between computational efficiency and enhancement performance. Our model demonstrates competitive results in both low-light image enhancement and defogging tasks, achieving a Peak Signal-to-Noise Ratio (PSNR) exceeding 20. Furthermore, it processes 1280x720 high-definition (HD) images at an impressive speed of 0.03 seconds per frame. This remarkable processing efficiency makes the model highly suitable for deployment in high-performance, resource-constrained systems, such as home surveillance equipment and autonomous vehicle platforms.

학위연월
2025년2월
지도교수
이도훈
키워드
Defogging, Low light enhancement, Lightweight, Real time Surveillance video, Low-cost devices
소개 웹페이지
https://dengbuqi.github.io/LMCVNET_Official_Page/
첨부파일
첨부파일이(가) 없습니다.
다음글
Toward Immersive Multiview Video Streaming through Viewpoint-centric Quality Adaptation
탄중 디온 2024-10-14 10:09:08.633
이전글
Enhancing Nested Entity Recognition Using Nested Rule-Based Method on Neural Network
양홍진 2024-10-09 19:06:25.42
RSS 2.0 123
게시물 검색
박사학위논문
번호 제목 작성자 작성일 첨부파일 조회수
123 Uncertainty-Based Hybrid Deep Learning Approach fo 멘가라 악셀 기드온 2024.12.10 0 62
122 Effective Deep Learning Primitives Design for Bina 황선진 2024.10.14 0 78
121 Toward Immersive Multiview Video Streaming through 탄중 디온 2024.10.14 0 44
120 A Low-cost Deep Learning Model for Real-time Low L 등 제강 2024.10.10 0 82
119 Enhancing Nested Entity Recognition Using Nested R 양홍진 2024.10.09 0 70
118 다양한 도메인과 데이터 형식에 강건한 사전학습 언어모델 기반의 표 질의응답 방법 조상현 2024.10.09 0 66
117 Trust Guard Extension for Enhanced Security Featur 김해용 2024.05.04 0 90
116 Task-Specific Differential Private Data Publish Me 신진명 2024.04.09 0 116
115 Advanced Defense Framework against Physical Advers 김용수 2024.04.08 0 137
114 한글 메신저 채팅의 크로스 텍스팅 탐지를 위한 저자 검증 모형 이다영 2024.04.05 0 110
113 상태 기반 테스트 시나리오 보강 방법 이선열 2023.10.17 0 190
112 Manufacturing Testing Automation FrameworkBased on 강효은 2023.10.17 0 203
111 Synthesizing Robust Physical Camouflage for Univer 수랸토 나우팔 2023.10.16 0 181
110 복잡도 다양성을 고려한 C 프로그램의 시험 용이성 예측 모형 구축 방법 최현재 2023.10.16 0 173
109 Design and Optimization of Quantum Arithmetic Circ 라라사티 하라스타 타티마 2023.10.13 0 206
108 Improving 6TiSCH Network Formation and Transmissio 파와즈 자키 자키얄 2023.10.10 0 184
107 저지연 고신뢰 운전자 프로파일링을 위한 딥러닝 모델 및 조기 종료 기법 임재봉 2023.10.08 0 268
106 802.11ax 대규모 Wi-Fi 환경의 심층 생성 모델을 활용한 트래픽 모델링 및 AP 이재민 2023.04.07 0 156
105 뉴런 클러스터를 활용한 합성곱 신경망 이미지 분류 신뢰성 향상 방법 이영우 2023.04.06 0 155
104 Trust Guard Extension Framework for Enhanced Secur 김해용 2023.04.06 0 120