Real-Time Long-Wave Infrared Semantic Segmentation With Adaptive Noise Reduction and Feature Fusion
Semantic segmentation in the Long-Wave Infrared (LWIR) domain is critical for a wide range of applications including emergency response, industrial safety monitoring, and public building security. However, LWIR images often suffer from inherent pattern noise (known as fixed-pattern noise) and ambigu...
Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IEEE
2025-01-01
|
| Series: | IEEE Access |
| Subjects: | |
| Online Access: | https://ieeexplore.ieee.org/document/10933928/ |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1849389193406447616 |
|---|---|
| author | Haejun Bae Dong-Goo Kang Minhye Chang Kye Young Jeong Byung Cheol Song |
| author_facet | Haejun Bae Dong-Goo Kang Minhye Chang Kye Young Jeong Byung Cheol Song |
| author_sort | Haejun Bae |
| collection | DOAJ |
| description | Semantic segmentation in the Long-Wave Infrared (LWIR) domain is critical for a wide range of applications including emergency response, industrial safety monitoring, and public building security. However, LWIR images often suffer from inherent pattern noise (known as fixed-pattern noise) and ambiguous object boundaries that hinder accurate segmentation. To address these challenges, this study presents a novel real-time semantic segmentation framework specifically designed for LWIR images. The framework incorporates a Stripe noise Denoising Module (SDM) and a Boundary Enhancement Module (BEM), which leverage frequency-domain filtering and learnable weights to adaptively process noisy input data and improve boundary prediction. In addition, a Multi-Stream Fusion Module (MSFM) integrates multi-scale semantic features with boundary information, thereby enhancing segmentation accuracy across diverse object scales. The proposed method demonstrates state-of-the-art performance in both accuracy and efficiency on multiple datasets, including KERInha and SODA. Extensive qualitative and quantitative evaluations further validate its robustness, particularly in scenarios where RGB imagery is unavailable. By eliminating the need for supplementary information such as depth data, this approach facilitates precise indoor and outdoor segmentation tasks with lightweight computation, making it highly suitable for real-world applications. Code is available at <uri>https://github.com/saha3jet/pytorch\_efficientirseg</uri>. |
| format | Article |
| id | doaj-art-52878fbf30d94edba920e60a4cb1e4e4 |
| institution | Kabale University |
| issn | 2169-3536 |
| language | English |
| publishDate | 2025-01-01 |
| publisher | IEEE |
| record_format | Article |
| series | IEEE Access |
| spelling | doaj-art-52878fbf30d94edba920e60a4cb1e4e42025-08-20T03:42:02ZengIEEEIEEE Access2169-35362025-01-0113519115192110.1109/ACCESS.2025.355278210933928Real-Time Long-Wave Infrared Semantic Segmentation With Adaptive Noise Reduction and Feature FusionHaejun Bae0https://orcid.org/0009-0007-8029-7391Dong-Goo Kang1https://orcid.org/0000-0002-4840-0025Minhye Chang2https://orcid.org/0000-0003-1460-6807Kye Young Jeong3Byung Cheol Song4https://orcid.org/0000-0001-8742-3433Department of Electrical and Computer Engineering, Inha University, Incheon, South KoreaKorea Electrotechnology Research Institute, Ansan, South KoreaKorea Electrotechnology Research Institute, Ansan, South KoreaKorea Electrotechnology Research Institute, Ansan, South KoreaDepartment of Electrical and Computer Engineering, Inha University, Incheon, South KoreaSemantic segmentation in the Long-Wave Infrared (LWIR) domain is critical for a wide range of applications including emergency response, industrial safety monitoring, and public building security. However, LWIR images often suffer from inherent pattern noise (known as fixed-pattern noise) and ambiguous object boundaries that hinder accurate segmentation. To address these challenges, this study presents a novel real-time semantic segmentation framework specifically designed for LWIR images. The framework incorporates a Stripe noise Denoising Module (SDM) and a Boundary Enhancement Module (BEM), which leverage frequency-domain filtering and learnable weights to adaptively process noisy input data and improve boundary prediction. In addition, a Multi-Stream Fusion Module (MSFM) integrates multi-scale semantic features with boundary information, thereby enhancing segmentation accuracy across diverse object scales. The proposed method demonstrates state-of-the-art performance in both accuracy and efficiency on multiple datasets, including KERInha and SODA. Extensive qualitative and quantitative evaluations further validate its robustness, particularly in scenarios where RGB imagery is unavailable. By eliminating the need for supplementary information such as depth data, this approach facilitates precise indoor and outdoor segmentation tasks with lightweight computation, making it highly suitable for real-world applications. Code is available at <uri>https://github.com/saha3jet/pytorch\_efficientirseg</uri>.https://ieeexplore.ieee.org/document/10933928/Long-wave infrared image segmentationnoise reductionfeature fusionboundary enhancement |
| spellingShingle | Haejun Bae Dong-Goo Kang Minhye Chang Kye Young Jeong Byung Cheol Song Real-Time Long-Wave Infrared Semantic Segmentation With Adaptive Noise Reduction and Feature Fusion IEEE Access Long-wave infrared image segmentation noise reduction feature fusion boundary enhancement |
| title | Real-Time Long-Wave Infrared Semantic Segmentation With Adaptive Noise Reduction and Feature Fusion |
| title_full | Real-Time Long-Wave Infrared Semantic Segmentation With Adaptive Noise Reduction and Feature Fusion |
| title_fullStr | Real-Time Long-Wave Infrared Semantic Segmentation With Adaptive Noise Reduction and Feature Fusion |
| title_full_unstemmed | Real-Time Long-Wave Infrared Semantic Segmentation With Adaptive Noise Reduction and Feature Fusion |
| title_short | Real-Time Long-Wave Infrared Semantic Segmentation With Adaptive Noise Reduction and Feature Fusion |
| title_sort | real time long wave infrared semantic segmentation with adaptive noise reduction and feature fusion |
| topic | Long-wave infrared image segmentation noise reduction feature fusion boundary enhancement |
| url | https://ieeexplore.ieee.org/document/10933928/ |
| work_keys_str_mv | AT haejunbae realtimelongwaveinfraredsemanticsegmentationwithadaptivenoisereductionandfeaturefusion AT donggookang realtimelongwaveinfraredsemanticsegmentationwithadaptivenoisereductionandfeaturefusion AT minhyechang realtimelongwaveinfraredsemanticsegmentationwithadaptivenoisereductionandfeaturefusion AT kyeyoungjeong realtimelongwaveinfraredsemanticsegmentationwithadaptivenoisereductionandfeaturefusion AT byungcheolsong realtimelongwaveinfraredsemanticsegmentationwithadaptivenoisereductionandfeaturefusion |