Blind HDR image quality assessment based on aggregating perception and inference features
Abstract High Dynamic Range (HDR) images, with their expanded range of brightness and color, provide a far more realistic and immersive viewing experience compared to Low Dynamic Range (LDR) images. However, the significant increase in peak luminance and contrast inherent in HDR images often accentu...
Saved in:
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Nature Portfolio
2025-03-01
|
| Series: | Scientific Reports |
| Subjects: | |
| Online Access: | https://doi.org/10.1038/s41598-025-94005-1 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1850063830763175936 |
|---|---|
| author | Donghui Wan Xiuhua Jiang Qiangguo Yu |
| author_facet | Donghui Wan Xiuhua Jiang Qiangguo Yu |
| author_sort | Donghui Wan |
| collection | DOAJ |
| description | Abstract High Dynamic Range (HDR) images, with their expanded range of brightness and color, provide a far more realistic and immersive viewing experience compared to Low Dynamic Range (LDR) images. However, the significant increase in peak luminance and contrast inherent in HDR images often accentuates artifacts, thus limiting the effectiveness of traditional LDR-based image quality assessment (IQA) algorithms when applied to HDR content. To address this, we propose a novel blind IQA method tailored specifically for HDR images, which incorporates both the perception and inference processes of the human visual system (HVS). Our approach begins with multi-scale Retinex decomposition to generate reflectance maps with varying sensitivity, followed by the calculation of gradient similarities from these maps to model the perception process. Deep feature maps are then extracted from the last pooling layer of a pretrained VGG16 network to capture inference characteristics. These gradient similarity maps and deep feature maps are subsequently aggregated for quality prediction using support vector regression (SVR). Experimental results demonstrate that the proposed method achieves outstanding performance, outperforming other state-of-the-art HDR IQA metrics. |
| format | Article |
| id | doaj-art-e920ec9761694f23abfb5f7c7cfbc7ee |
| institution | DOAJ |
| issn | 2045-2322 |
| language | English |
| publishDate | 2025-03-01 |
| publisher | Nature Portfolio |
| record_format | Article |
| series | Scientific Reports |
| spelling | doaj-art-e920ec9761694f23abfb5f7c7cfbc7ee2025-08-20T02:49:29ZengNature PortfolioScientific Reports2045-23222025-03-0115111210.1038/s41598-025-94005-1Blind HDR image quality assessment based on aggregating perception and inference featuresDonghui Wan0Xiuhua Jiang1Qiangguo Yu2State Key Laboratory of Media Convergence and Communication, Communication University of ChinaState Key Laboratory of Media Convergence and Communication, Communication University of ChinaSchool of Electronic Information, Huzhou CollegeAbstract High Dynamic Range (HDR) images, with their expanded range of brightness and color, provide a far more realistic and immersive viewing experience compared to Low Dynamic Range (LDR) images. However, the significant increase in peak luminance and contrast inherent in HDR images often accentuates artifacts, thus limiting the effectiveness of traditional LDR-based image quality assessment (IQA) algorithms when applied to HDR content. To address this, we propose a novel blind IQA method tailored specifically for HDR images, which incorporates both the perception and inference processes of the human visual system (HVS). Our approach begins with multi-scale Retinex decomposition to generate reflectance maps with varying sensitivity, followed by the calculation of gradient similarities from these maps to model the perception process. Deep feature maps are then extracted from the last pooling layer of a pretrained VGG16 network to capture inference characteristics. These gradient similarity maps and deep feature maps are subsequently aggregated for quality prediction using support vector regression (SVR). Experimental results demonstrate that the proposed method achieves outstanding performance, outperforming other state-of-the-art HDR IQA metrics.https://doi.org/10.1038/s41598-025-94005-1High dynamic range imageNo reference quality assessmentPerception processInference process |
| spellingShingle | Donghui Wan Xiuhua Jiang Qiangguo Yu Blind HDR image quality assessment based on aggregating perception and inference features Scientific Reports High dynamic range image No reference quality assessment Perception process Inference process |
| title | Blind HDR image quality assessment based on aggregating perception and inference features |
| title_full | Blind HDR image quality assessment based on aggregating perception and inference features |
| title_fullStr | Blind HDR image quality assessment based on aggregating perception and inference features |
| title_full_unstemmed | Blind HDR image quality assessment based on aggregating perception and inference features |
| title_short | Blind HDR image quality assessment based on aggregating perception and inference features |
| title_sort | blind hdr image quality assessment based on aggregating perception and inference features |
| topic | High dynamic range image No reference quality assessment Perception process Inference process |
| url | https://doi.org/10.1038/s41598-025-94005-1 |
| work_keys_str_mv | AT donghuiwan blindhdrimagequalityassessmentbasedonaggregatingperceptionandinferencefeatures AT xiuhuajiang blindhdrimagequalityassessmentbasedonaggregatingperceptionandinferencefeatures AT qiangguoyu blindhdrimagequalityassessmentbasedonaggregatingperceptionandinferencefeatures |