From Fog to Failure: The Unintended Consequences of Dehazing on Object Detection in Clear Images
This study explores the challenges of integrating human visual cue-based dehazing into object detection, given the selective nature of human perception. While human vision adapts dynamically to environmental conditions, computational dehazing does not always enhance detection uniformly. We propose a...
Guardado en:
| Autores principales: | , |
|---|---|
| Formato: | Artículo |
| Publicado: |
2025
|
| Materias: | |
| Acceso en línea: | https://arxiv.org/abs/2502.02027 http://arxiv.org/abs/2502.02027 |
| Etiquetas: |
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1879176795101593600 |
|---|---|
| author | Kumar, Ashutosh Chadha, Aman |
| author_browse | Chadha, Aman Kumar, Ashutosh |
| author_facet | Kumar, Ashutosh Chadha, Aman |
| author_sort | Kumar, Ashutosh |
| building | arXiv |
| collection | Acceso abierto |
| description | This study explores the challenges of integrating human visual cue-based dehazing into object detection, given the selective nature of human perception. While human vision adapts dynamically to environmental conditions, computational dehazing does not always enhance detection uniformly. We propose a multi-stage framework where a lightweight detector identifies regions of interest (RoIs), which are then improved via spatial attention-based dehazing before final detection by a heavier model. Though effective in foggy conditions, this approach unexpectedly degrades the performance on clear images. We analyze this phenomenon, investigate possible causes, and offer insights for designing hybrid pipelines that balance enhancement and detection. Our findings highlight the need for selective preprocessing and challenge assumptions about universal benefits from cascading transformations. |
| first_indexed | 2026-10-16T00:50:40Z |
| format | Article |
| fullrecord | <oai_dc:dc xmlns:oai_dc="http://www.openarchives.org/OAI/2.0/oai_dc/" xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:schemaLocation="http://www.openarchives.org/OAI/2.0/oai_dc/ https://www.openarchives.org/OAI/2.0/oai_dc.xsd"><identifier>arxiv-2502.02027</identifier><datestamp>2025-03-18</datestamp>
<dc:title>From Fog to Failure: The Unintended Consequences of Dehazing on Object Detection in Clear Images</dc:title>
<dc:creator>Kumar, Ashutosh</dc:creator>
<dc:creator>Chadha, Aman</dc:creator>
<dc:subject>Computer Vision and Pattern Recognition</dc:subject>
<dc:subject>Artificial Intelligence</dc:subject>
<dc:description>This study explores the challenges of integrating human visual cue-based dehazing into object detection, given the selective nature of human perception. While human vision adapts dynamically to environmental conditions, computational dehazing does not always enhance detection uniformly. We propose a multi-stage framework where a lightweight detector identifies regions of interest (RoIs), which are then improved via spatial attention-based dehazing before final detection by a heavier model. Though effective in foggy conditions, this approach unexpectedly degrades the performance on clear images. We analyze this phenomenon, investigate possible causes, and offer insights for designing hybrid pipelines that balance enhancement and detection. Our findings highlight the need for selective preprocessing and challenge assumptions about universal benefits from cascading transformations. </dc:description>
<dc:date>2025-02-04</dc:date>
<dc:date>2025-03-16</dc:date>
<dc:type>text</dc:type>
<dc:identifier>https://arxiv.org/abs/2502.02027</dc:identifier>
<dc:identifier>http://arxiv.org/abs/2502.02027</dc:identifier>
</oai_dc:dc> |
| id | arxiv-2502.02027 |
| institution | arXiv |
| last_indexed | 2026-10-16T00:50:40Z |
| publishDate | 2025 |
| publishDateRange | 2025 |
| publishDateSort | 2025 |
| record_format | oai_dc |
| spelling | arxiv-2502.020272025-03-18 From Fog to Failure: The Unintended Consequences of Dehazing on Object Detection in Clear Images Kumar, Ashutosh Chadha, Aman Computer Vision and Pattern Recognition Artificial Intelligence This study explores the challenges of integrating human visual cue-based dehazing into object detection, given the selective nature of human perception. While human vision adapts dynamically to environmental conditions, computational dehazing does not always enhance detection uniformly. We propose a multi-stage framework where a lightweight detector identifies regions of interest (RoIs), which are then improved via spatial attention-based dehazing before final detection by a heavier model. Though effective in foggy conditions, this approach unexpectedly degrades the performance on clear images. We analyze this phenomenon, investigate possible causes, and offer insights for designing hybrid pipelines that balance enhancement and detection. Our findings highlight the need for selective preprocessing and challenge assumptions about universal benefits from cascading transformations. 2025-02-04 2025-03-16 text https://arxiv.org/abs/2502.02027 http://arxiv.org/abs/2502.02027 |
| spellingShingle | Computer Vision and Pattern Recognition Artificial Intelligence Kumar, Ashutosh Chadha, Aman From Fog to Failure: The Unintended Consequences of Dehazing on Object Detection in Clear Images |
| title | From Fog to Failure: The Unintended Consequences of Dehazing on Object Detection in Clear Images |
| title_full | From Fog to Failure: The Unintended Consequences of Dehazing on Object Detection in Clear Images |
| title_fullStr | From Fog to Failure: The Unintended Consequences of Dehazing on Object Detection in Clear Images |
| title_full_unstemmed | From Fog to Failure: The Unintended Consequences of Dehazing on Object Detection in Clear Images |
| title_short | From Fog to Failure: The Unintended Consequences of Dehazing on Object Detection in Clear Images |
| title_sort | from fog to failure the unintended consequences of dehazing on object detection in clear images |
| topic | Computer Vision and Pattern Recognition Artificial Intelligence |
| url | https://arxiv.org/abs/2502.02027 http://arxiv.org/abs/2502.02027 |
| work_keys_str_mv | AT kumarashutosh fromfogtofailuretheunintendedconsequencesofdehazingonobjectdetectioninclearimages AT chadhaaman fromfogtofailuretheunintendedconsequencesofdehazingonobjectdetectioninclearimages |