{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,9,23]],"date-time":"2026-09-23T07:01:50Z","timestamp":1790146910397,"version":"4.0.1"},"reference-count":46,"publisher":"Frontiers Media SA","license":[{"start":{"date-parts":[[2026,9,23]],"date-time":"2026-09-23T00:00:00Z","timestamp":1790121600000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"content-domain":{"domain":["frontiersin.org"],"crossmark-restriction":true},"short-container-title":["Front. Artif. Intell."],"abstract":"<jats:p>\n                    Image dehazing is a basic low-level vision task designed to recover high-fidelity images from hazed images. In autonomous driving, aerial surveillance, and outdoor scene recognition, image quality is significantly degraded by light attenuation due to scattering in the propagation medium, especially under hazy, foggy, or smoggy conditions. While remarkable advances have been made with existing deep learning methods, they remain constrained by over-smoothing, ineffective use of haze frequency, and insufficient consideration of multi-scale feature interactions. To meet these challenges, we introduce a novel single-image dehazing framework, called Low-Frequency Enhanced Feature Transmission Network (LEFTNet), which consists of four special modules: the Haze Suppression Block (HSB), the Dual-Branch Block (DBB), the Cross-Scale Fusion Block (CSFB), and the Channel Mining Mechanism (CMM). In the first stage, a hazy input image is decomposed into two complementary low-frequency representations: a haze-related component (\n                    <jats:inline-formula>\n                      <mml:math xmlns:mml=\"http:\/\/www.w3.org\/1998\/Math\/MathML\" id=\"M1\">\n                        <mml:msubsup>\n                          <mml:mrow>\n                            <mml:mi>I<\/mml:mi>\n                          <\/mml:mrow>\n                          <mml:mrow>\n                            <mml:mstyle class=\"text\">\n                              <mml:mtext class=\"textrm\" mathvariant=\"normal\">haze<\/mml:mtext>\n                            <\/mml:mstyle>\n                          <\/mml:mrow>\n                          <mml:mrow>\n                            <mml:mi>L<\/mml:mi>\n                          <\/mml:mrow>\n                        <\/mml:msubsup>\n                      <\/mml:math>\n                    <\/jats:inline-formula>\n                    ) and a frequency-domain component (\n                    <jats:inline-formula>\n                      <mml:math xmlns:mml=\"http:\/\/www.w3.org\/1998\/Math\/MathML\" id=\"M2\">\n                        <mml:msubsup>\n                          <mml:mrow>\n                            <mml:mi>I<\/mml:mi>\n                          <\/mml:mrow>\n                          <mml:mrow>\n                            <mml:mstyle class=\"text\">\n                              <mml:mtext class=\"textrm\" mathvariant=\"normal\">freq<\/mml:mtext>\n                            <\/mml:mstyle>\n                          <\/mml:mrow>\n                          <mml:mrow>\n                            <mml:mi>L<\/mml:mi>\n                          <\/mml:mrow>\n                        <\/mml:msubsup>\n                      <\/mml:math>\n                    <\/jats:inline-formula>\n                    ), which are jointly processed by the HSB for early, channel-selective haze attenuation. Then, the encoder extracts multi-scale hierarchical features, which are cleaned up at each stage by the Channel Mining module. The DBB's residual refinement enables intra-image global-to-local feature interactions, while the CSFB's learnable fusion coefficients enable inter-scale feature fusion. A dual fully connected bottleneck with Global Average Pooling (GAP) suppresses haze-dominant channels while enhancing structure-sensitive channels. Two symmetric branches ensure complementary representation learning, and a hybrid loss function comprising L1 loss and perceptual LSSIM loss assures both pixel-wise accuracy and perceptual fidelity. Comprehensive experiments on four benchmarks, RESIDE-6K, I-Haze, O-Haze, and NH-Haze, demonstrate that LEFTNet achieves competitive performance across the evaluated Peak Signal-to-Noise Ratio (PSNR), Structural Similarity Index Measure (SSIM), Average Mean Brightness Error (AMBE), Gradient Magnitude Similarity Deviation (GMSD), and Visual Saliency-Induced Index (VSI) metrics compared with recent dehazing methods.\n                  <\/jats:p>","DOI":"10.3389\/frai.2026.1938382","type":"journal-article","created":{"date-parts":[[2026,9,23]],"date-time":"2026-09-23T06:20:19Z","timestamp":1790144419000},"update-policy":"https:\/\/doi.org\/10.3389\/crossmark-policy","source":"Crossref","is-referenced-by-count":0,"title":["LEFTNet: low-frequency enhanced feature transmission network for dual-branch collaborative single image dehazing"],"prefix":"10.3389","volume":"9","author":[{"given":"Beesetti","family":"Vanitha","sequence":"first","affiliation":[{"name":"School of Computer Science and Engineering, VIT-AP University, Amaravati","place":["Andhra Pradesh, India"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Chintakindi","family":"Balaram Murthy","sequence":"additional","affiliation":[{"name":"School of Computer Science and Engineering, VIT-AP University, Amaravati","place":["Andhra Pradesh, India"]}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"1965","published-online":{"date-parts":[[2026,9,23]]},"reference":[{"key":"B1","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-030-01449-0_52","article-title":"\u201cI-HAZE: a dehazing benchmark with real hazy and haze-free indoor images,\u201d","author":"Ancuti","year":"2018","journal-title":"Advanced concepts for intelligent vision systems"},{"key":"B2","doi-asserted-by":"publisher","DOI":"10.1109\/CVPRW50498.2020.00230","article-title":"\u201cNH-HAZE: an image dehazing benchmark with non-homogeneous hazy and haze-free images,\u201d","author":"Ancuti","year":"2020","journal-title":"2020 IEEE\/CVF conference on computer vision and pattern recognition workshops (CVPRW)"},{"key":"B3","doi-asserted-by":"publisher","DOI":"10.1109\/CVPRW.2018.00119","article-title":"\u201cO-HAZE: a dehazing benchmark with real hazy and haze-free outdoor images,\u201d","author":"Ancuti","year":"2018","journal-title":"2018 IEEE\/CVF conference on computer vision and pattern recognition workshops (CVPRW)"},{"key":"B4","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2016.185","article-title":"\u201cNon-local image dehazing,\u201d","author":"Berman","year":"2016","journal-title":"2016 IEEE conference on computer vision and pattern recognition (CVPR)"},{"key":"B5","doi-asserted-by":"publisher","first-page":"5187","DOI":"10.1109\/TIP.2016.2598681","article-title":"DehazeNet: an end-to-end system for single image haze removal","volume":"25","author":"Cai","year":"2016","journal-title":"IEEE Trans. Image Proc"},{"key":"B6","doi-asserted-by":"publisher","first-page":"1002","DOI":"10.1109\/TIP.2024.3354108","article-title":"DEA-Net: single image dehazing based on detail-enhanced convolution and content-guided attention","volume":"33","author":"Chen","year":"2024","journal-title":"IEEE Trans. Image Proc"},{"key":"B7","doi-asserted-by":"publisher","first-page":"2154","DOI":"10.1109\/CVPR42600.2020.00223","article-title":"\u201cMulti-scale boosted dehazing network with dense feature fusion,\u201d","author":"Dong","year":"2020","journal-title":"2020 IEEE\/CVF conference on computer vision and pattern recognition (CVPR)"},{"key":"B8","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1145\/1360612.1360671","article-title":"Single image dehazing","volume":"27","author":"Fattal","year":"2008","journal-title":"ACM Trans. Graph"},{"key":"B9","doi-asserted-by":"publisher","first-page":"135","DOI":"10.1016\/j.sigpro.2018.03.008","article-title":"Image dehazing by artificial multiple-exposure image fusion","volume":"149","author":"Galdran","year":"2018","journal-title":"Signal Proc"},{"key":"B10","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR52688.2022.00572","article-title":"\u201cImage dehazing transformer with transmission-aware 3D position embedding,\u201d","author":"Guo","year":"2022","journal-title":"2022 IEEE\/CVF conference on computer vision and pattern recognition (CVPR)"},{"key":"B11","doi-asserted-by":"publisher","DOI":"10.1109\/ICPR.2010.579","article-title":"\u201cImage quality metrics: PSNR vs. SSIM,\u201d","author":"Hore","year":"2010","journal-title":"2010 20th international conference on pattern recognition"},{"key":"B12","article-title":"WaveDH: wavelet sub-bands guided ConvNet for efficient image dehazing","author":"Hwang","year":"2024","journal-title":"arXiv preprint arXiv:2404.01604"},{"key":"B13","doi-asserted-by":"publisher","first-page":"2341","DOI":"10.1109\/TPAMI.2010.168","article-title":"Single image haze removal using dark channel prior","volume":"33","author":"Kaiming","year":"2011","journal-title":"IEEE Trans. Pattern Anal. Mach. Intell"},{"key":"B14","doi-asserted-by":"publisher","DOI":"10.1109\/WACV56688.2023.00624","article-title":"\u201cAerial image dehazing with attentive deformable transformers,\u201d","author":"Kulkarni","year":"2023","journal-title":"2023 IEEE\/CVF winter conference on applications of computer vision (WACV)"},{"key":"B15","doi-asserted-by":"publisher","first-page":"4780","DOI":"10.1109\/ICCV.2017.511","article-title":"\u201cAOD-net: all-in-one dehazing network,\u201d","author":"Li","year":"2017","journal-title":"2017 IEEE international conference on computer vision (ICCV)"},{"key":"B16","doi-asserted-by":"publisher","first-page":"492","DOI":"10.1109\/TIP.2018.2867951","article-title":"Benchmarking single-image dehazing and beyond","volume":"28","author":"Li","year":"2019","journal-title":"IEEE Trans. Image Proc"},{"key":"B17","doi-asserted-by":"publisher","DOI":"10.1109\/ICPR48806.2021.9412595","article-title":"\u201cFast region-adaptive defogging and enhancement for outdoor images containing sky,\u201d","author":"Li","year":"2021","journal-title":"2020 25th international conference on pattern recognition (ICPR)"},{"key":"B18","doi-asserted-by":"publisher","first-page":"3934","DOI":"10.1109\/TMM.2021.3110483","article-title":"Joint contrast enhancement and exposure fusion for real-world image dehazing","volume":"24","author":"Liu","year":"2022","journal-title":"IEEE Trans. Multim"},{"key":"B19","doi-asserted-by":"publisher","first-page":"870","DOI":"10.1109\/TITS.2022.3210455","article-title":"GridDehazeNet+: an enhanced multi-scale network with intra-task knowledge transfer for single image dehazing","volume":"24","author":"Liu","year":"2023","journal-title":"IEEE Trans. Intell. Transport. Syst"},{"key":"B20","volume-title":"Optics of the Atmosphere: Scattering by Molecules and Particles","author":"McCartney","year":"1976"},{"key":"B21","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2000.855874","article-title":"\u201cChromatic framework for vision in bad weather,\u201d","author":"Narasimhan","year":"2000","journal-title":"Proceedings IEEE conference on computer vision and pattern recognition. CVPR 2000 (Cat. No.PR00662)"},{"key":"B22","doi-asserted-by":"publisher","first-page":"11875","DOI":"10.1038\/s41598-025-92061-1","article-title":"SAD-Net: a full spectral self-attention detail enhancement network for single image dehazing","volume":"15","author":"Niu","year":"2025","journal-title":"Sci. Rep"},{"key":"B23","doi-asserted-by":"publisher","first-page":"281","DOI":"10.1007\/s00034-024-02837-5","article-title":"SWAM-Net+: selective wavelet attentive M-network+ for single image dehazing","volume":"44","author":"Nuthi","year":"2025","journal-title":"Circ. Syst. Signal Proc"},{"key":"B24","doi-asserted-by":"publisher","DOI":"10.1609\/aaai.v34i07.6865","article-title":"\u201cFFA-Net: feature fusion attention network for single image dehazing,\u201d","author":"Qin","year":"2020","journal-title":"Proceedings of the AAAI conference on artificial intelligence"},{"key":"B25","doi-asserted-by":"publisher","first-page":"3522","DOI":"10.1109\/TIP.2015.2446191","article-title":"A fast single image haze removal algorithm using color attenuation prior","volume":"24","author":"Qingsong","year":"2015","journal-title":"IEEE Trans. Image Proc"},{"key":"B26","doi-asserted-by":"publisher","DOI":"10.1109\/ICCV51070.2023.01176","article-title":"\u201cMB-TaylorFormer: multi-branch efficient transformer expanded by Taylor formula for image dehazing,\u201d","author":"Qiu","year":"2023","journal-title":"2023 IEEE\/CVF international conference on computer vision (ICCV)"},{"key":"B27","doi-asserted-by":"publisher","first-page":"104710","DOI":"10.1016\/j.dsp.2024.104710","article-title":"TSMD-Net: a two-stage mixed dehazing network with feature fusion and multi-window self attention on Jetson board","volume":"155","author":"Raju","year":"2024","journal-title":"Digit. Signal Process"},{"key":"B28","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-319-46475-6_10","article-title":"\u201cSingle image dehazing via multi-scale convolutional neural networks,\u201d","author":"Ren","year":"2016","journal-title":"Computer vision"},{"key":"B29","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2018.00343","article-title":"\u201cGated fusion network for single image dehazing,\u201d","author":"Ren","year":"2018","journal-title":"2018 IEEE\/CVF conference on computer vision and pattern recognition"},{"key":"B30","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR42600.2020.00288","article-title":"\u201cDomain adaptation for image dehazing,\u201d","author":"Shao","year":"2020","journal-title":"2020 IEEE\/CVF conference on computer vision and pattern recognition (CVPR)"},{"key":"B31","doi-asserted-by":"publisher","first-page":"1927","DOI":"10.1109\/TIP.2023.3256763","article-title":"Vision transformers for single image dehazing","volume":"32","author":"Song","year":"2023","journal-title":"IEEE Trans. Image Proc"},{"key":"B32","doi-asserted-by":"publisher","first-page":"1310","DOI":"10.1109\/TCE.2003.1261234","article-title":"Minimum mean brightness error bi-histogram equalization in contrast enhancement","volume":"49","year":"2003","journal-title":"IEEE Trans. Consumer Electr"},{"key":"B33","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2008.4587643","article-title":"\u201cVisibility in bad weather from a single image,\u201d","author":"Tan","year":"2008","journal-title":"2008 IEEE conference on computer vision and pattern recognition"},{"key":"B34","doi-asserted-by":"publisher","first-page":"8968","DOI":"10.1109\/TIP.2021.3116790","article-title":"Light-DehazeNet: a novel lightweight CNN architecture for single image dehazing","volume":"30","author":"Ullah","year":"2021","journal-title":"IEEE Trans. Image Proc"},{"key":"B35","doi-asserted-by":"publisher","first-page":"110397","DOI":"10.1016\/j.patcog.2024.110397","article-title":"Dual-path dehazing network with spatial-frequency feature fusion","volume":"151","author":"Wang","year":"2024","journal-title":"Pattern Recognit"},{"key":"B36","doi-asserted-by":"publisher","first-page":"684","DOI":"10.1109\/TIP.2013.2293423","article-title":"Gradient magnitude similarity deviation: a highly efficient perceptual image quality index","volume":"23","author":"Xue","year":"2014","journal-title":"IEEE Trans. Image Proc"},{"key":"B37","doi-asserted-by":"publisher","first-page":"106165","DOI":"10.1016\/j.neunet.2024.106165","article-title":"Priors-assisted dehazing network with attention supervision and detail preservation","volume":"173","author":"Yi","year":"2024","journal-title":"Neural Netw"},{"key":"B38","article-title":"High-quality image dehazing with diffusion model","author":"Yu","year":"2023","journal-title":"arXiv preprint arXiv:2308.11949"},{"key":"B39","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-031-19800-7_11","article-title":"\u201cFrequency and spatial dual guidance for image dehazing,\u201d","author":"Yu","year":"2022","journal-title":"Computer vision"},{"key":"B40","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2018.00337","article-title":"\u201cDensely connected pyramid dehazing network,\u201d","author":"Zhang","year":"2018","journal-title":"2018 IEEE\/CVF conference on computer vision and pattern recognition"},{"key":"B41","doi-asserted-by":"publisher","first-page":"4270","DOI":"10.1109\/TIP.2014.2346028","article-title":"VSI: a visual saliency-induced index for perceptual image quality assessment","volume":"23","author":"Zhang","year":"2014","journal-title":"IEEE Trans. Image Proc"},{"key":"B42","article-title":"Hybrid local-global transformer for image dehazing","author":"Zhao","year":"2021","journal-title":"arXiv preprint arXiv:2109.07100"},{"key":"B43","doi-asserted-by":"publisher","first-page":"125803","DOI":"10.1016\/j.eswa.2024.125803","article-title":"Low-light stereo image enhancement and de-noising in the low-frequency information enhanced image space","volume":"265","author":"Zhao","year":"2025","journal-title":"Expert Syst. Appl"},{"key":"B44","article-title":"RSDehamba: lightweight vision mamba for remote sensing satellite image dehazing","author":"Zhou","year":"2024","journal-title":"arXiv preprint arXiv:2405.10030"},{"key":"B45","doi-asserted-by":"publisher","first-page":"600","DOI":"10.1109\/TIP.2003.819861","article-title":"Image quality assessment: from error visibility to structural similarity","volume":"13","author":"Zhou","year":"2004","journal-title":"IEEE Trans. Image Proc"},{"key":"B46","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1109\/TIM.2020.3024335","article-title":"A novel fast single image dehazing algorithm based on artificial multiexposure image fusion","volume":"70","author":"Zhu","year":"2021","journal-title":"IEEE Trans. Instrum. Meas"}],"container-title":["Frontiers in Artificial Intelligence"],"original-title":[],"link":[{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/frai.2026.1938382\/full","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,9,23]],"date-time":"2026-09-23T06:20:20Z","timestamp":1790144420000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/frai.2026.1938382\/full"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,9,23]]},"references-count":46,"alternative-id":["10.3389\/frai.2026.1938382"],"URL":"https:\/\/doi.org\/10.3389\/frai.2026.1938382","relation":{},"ISSN":["2624-8212"],"issn-type":[{"value":"2624-8212","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,9,23]]},"article-number":"1938382"}}