QMRNet: Quality Metric Regression for EO Image Quality Assessment and Super-Resolution
The latest advances in super-resolution have been tested with general-purpose images such as faces, landscapes and objects, but mainly unused for the task of super-resolving earth observation images. In this research paper, we benchmark state-of-the-art SR algorithms for distinct EO datasets using b...
| Main Authors: | , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2023-05-01
|
| Series: | Remote Sensing |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2072-4292/15/9/2451 |
| _version_ | 1797601710832615424 |
|---|---|
| author | David Berga Pau Gallés Katalin Takáts Eva Mohedano Laura Riordan-Chen Clara Garcia-Moll David Vilaseca Javier Marín |
| author_facet | David Berga Pau Gallés Katalin Takáts Eva Mohedano Laura Riordan-Chen Clara Garcia-Moll David Vilaseca Javier Marín |
| author_sort | David Berga |
| collection | DOAJ |
| description | The latest advances in super-resolution have been tested with general-purpose images such as faces, landscapes and objects, but mainly unused for the task of super-resolving earth observation images. In this research paper, we benchmark state-of-the-art SR algorithms for distinct EO datasets using both full-reference and no-reference image quality assessment metrics. We also propose a novel Quality Metric Regression Network (QMRNet) that is able to predict the quality (as a no-reference metric) by training on any property of the image (e.g., its resolution, its distortions, etc.) and also able to optimize SR algorithms for a specific metric objective. This work is part of the implementation of the framework IQUAFLOW, which has been developed for the evaluation of image quality and the detection and classification of objects as well as image compression in EO use cases. We integrated our experimentation and tested our QMRNet algorithm on predicting features such as blur, sharpness, snr, rer and ground sampling distance and obtained validation medRs below 1.0 (out of <i>N</i> = 50) and recall rates above 95%. The overall benchmark shows promising results for LIIF, CAR and MSRN and also the potential use of QMRNet as a loss for optimizing SR predictions. Due to its simplicity, QMRNet could also be used for other use cases and image domains, as its architecture and data processing is fully scalable. |
| first_indexed | 2024-03-11T04:07:32Z |
| format | Article |
| id | doaj.art-050e6af98eec4ae3b7ef7c2023591275 |
| institution | Directory Open Access Journal |
| issn | 2072-4292 |
| language | English |
| last_indexed | 2024-03-11T04:07:32Z |
| publishDate | 2023-05-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Remote Sensing |
| spelling | doaj.art-050e6af98eec4ae3b7ef7c20235912752023-11-17T23:40:22ZengMDPI AGRemote Sensing2072-42922023-05-01159245110.3390/rs15092451QMRNet: Quality Metric Regression for EO Image Quality Assessment and Super-ResolutionDavid Berga0Pau Gallés1Katalin Takáts2Eva Mohedano3Laura Riordan-Chen4Clara Garcia-Moll5David Vilaseca6Javier Marín7Eurecat, Centre Tecnològic de Catalunya, Tecnologies Multimèdia, 08005 Barcelona, SpainSatellogic Inc., Davidson, NC 28036, USASatellogic Inc., Davidson, NC 28036, USASatellogic Inc., Davidson, NC 28036, USASatellogic Inc., Davidson, NC 28036, USASatellogic Inc., Davidson, NC 28036, USASatellogic Inc., Davidson, NC 28036, USASatellogic Inc., Davidson, NC 28036, USAThe latest advances in super-resolution have been tested with general-purpose images such as faces, landscapes and objects, but mainly unused for the task of super-resolving earth observation images. In this research paper, we benchmark state-of-the-art SR algorithms for distinct EO datasets using both full-reference and no-reference image quality assessment metrics. We also propose a novel Quality Metric Regression Network (QMRNet) that is able to predict the quality (as a no-reference metric) by training on any property of the image (e.g., its resolution, its distortions, etc.) and also able to optimize SR algorithms for a specific metric objective. This work is part of the implementation of the framework IQUAFLOW, which has been developed for the evaluation of image quality and the detection and classification of objects as well as image compression in EO use cases. We integrated our experimentation and tested our QMRNet algorithm on predicting features such as blur, sharpness, snr, rer and ground sampling distance and obtained validation medRs below 1.0 (out of <i>N</i> = 50) and recall rates above 95%. The overall benchmark shows promising results for LIIF, CAR and MSRN and also the potential use of QMRNet as a loss for optimizing SR predictions. Due to its simplicity, QMRNet could also be used for other use cases and image domains, as its architecture and data processing is fully scalable.https://www.mdpi.com/2072-4292/15/9/2451super-resolutionquality assessmentbenchmarkdenoisingregressionautoencoder networks |
| spellingShingle | David Berga Pau Gallés Katalin Takáts Eva Mohedano Laura Riordan-Chen Clara Garcia-Moll David Vilaseca Javier Marín QMRNet: Quality Metric Regression for EO Image Quality Assessment and Super-Resolution Remote Sensing super-resolution quality assessment benchmark denoising regression autoencoder networks |
| title | QMRNet: Quality Metric Regression for EO Image Quality Assessment and Super-Resolution |
| title_full | QMRNet: Quality Metric Regression for EO Image Quality Assessment and Super-Resolution |
| title_fullStr | QMRNet: Quality Metric Regression for EO Image Quality Assessment and Super-Resolution |
| title_full_unstemmed | QMRNet: Quality Metric Regression for EO Image Quality Assessment and Super-Resolution |
| title_short | QMRNet: Quality Metric Regression for EO Image Quality Assessment and Super-Resolution |
| title_sort | qmrnet quality metric regression for eo image quality assessment and super resolution |
| topic | super-resolution quality assessment benchmark denoising regression autoencoder networks |
| url | https://www.mdpi.com/2072-4292/15/9/2451 |
| work_keys_str_mv | AT davidberga qmrnetqualitymetricregressionforeoimagequalityassessmentandsuperresolution AT paugalles qmrnetqualitymetricregressionforeoimagequalityassessmentandsuperresolution AT katalintakats qmrnetqualitymetricregressionforeoimagequalityassessmentandsuperresolution AT evamohedano qmrnetqualitymetricregressionforeoimagequalityassessmentandsuperresolution AT laurariordanchen qmrnetqualitymetricregressionforeoimagequalityassessmentandsuperresolution AT claragarciamoll qmrnetqualitymetricregressionforeoimagequalityassessmentandsuperresolution AT davidvilaseca qmrnetqualitymetricregressionforeoimagequalityassessmentandsuperresolution AT javiermarin qmrnetqualitymetricregressionforeoimagequalityassessmentandsuperresolution |