Remote Sensing Image Compression Based on the Multiple Prior Information
Learned image compression has achieved a series of breakthroughs for nature images, but there is little literature focusing on high-resolution remote sensing image (HRRSI) datasets. This paper focuses on designing a learned lossy image compression framework for compressing HRRSIs. Considering the lo...
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2023-04-01
|
| Series: | Remote Sensing |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2072-4292/15/8/2211 |
| _version_ | 1797603545547014144 |
|---|---|
| author | Chuan Fu Bo Du |
| author_facet | Chuan Fu Bo Du |
| author_sort | Chuan Fu |
| collection | DOAJ |
| description | Learned image compression has achieved a series of breakthroughs for nature images, but there is little literature focusing on high-resolution remote sensing image (HRRSI) datasets. This paper focuses on designing a learned lossy image compression framework for compressing HRRSIs. Considering the local and non-local redundancy contained in HRRSI, a mixed hyperprior network is designed to explore both the local and non-local redundancy in order to improve the accuracy of entropy estimation. In detail, a transformer-based hyperprior and a CNN-based hyperprior are fused for entropy estimation. Furthermore, to reduce the mismatch between training and testing, a three-stage training strategy is introduced to refine the network. In this training strategy, the entire network is first trained, and then some sub-networks are fixed while the others are trained. To evaluate the effectiveness of the proposed compression algorithm, the experiments are conducted on an HRRSI dataset. The results show that the proposed algorithm achieves comparable or better compression performance than some traditional and learned image compression algorithms, such as Joint Photographic Experts Group (JPEG) and JPEG2000. At a similar or lower bitrate, the proposed algorithm is about 2 dB higher than the PSNR value of JPEG2000. |
| first_indexed | 2024-03-11T04:33:38Z |
| format | Article |
| id | doaj.art-0ba609f884b645d3a868181f4f2e82fc |
| institution | Directory Open Access Journal |
| issn | 2072-4292 |
| language | English |
| last_indexed | 2024-03-11T04:33:38Z |
| publishDate | 2023-04-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Remote Sensing |
| spelling | doaj.art-0ba609f884b645d3a868181f4f2e82fc2023-11-17T21:13:33ZengMDPI AGRemote Sensing2072-42922023-04-01158221110.3390/rs15082211Remote Sensing Image Compression Based on the Multiple Prior InformationChuan Fu0Bo Du1The State Key Laboratory of Information Engineering in Surveying, Mapping and Remote Sensing, Wuhan University, Wuhan 430079, ChinaThe National Engineering Research Center for Multimedia Software, Institute of Artificial Intelligence, School of Computer Science, and Hubei Key Laboratory of Multimedia and Network Communication Engineering, Wuhan University, Wuhan 430079, ChinaLearned image compression has achieved a series of breakthroughs for nature images, but there is little literature focusing on high-resolution remote sensing image (HRRSI) datasets. This paper focuses on designing a learned lossy image compression framework for compressing HRRSIs. Considering the local and non-local redundancy contained in HRRSI, a mixed hyperprior network is designed to explore both the local and non-local redundancy in order to improve the accuracy of entropy estimation. In detail, a transformer-based hyperprior and a CNN-based hyperprior are fused for entropy estimation. Furthermore, to reduce the mismatch between training and testing, a three-stage training strategy is introduced to refine the network. In this training strategy, the entire network is first trained, and then some sub-networks are fixed while the others are trained. To evaluate the effectiveness of the proposed compression algorithm, the experiments are conducted on an HRRSI dataset. The results show that the proposed algorithm achieves comparable or better compression performance than some traditional and learned image compression algorithms, such as Joint Photographic Experts Group (JPEG) and JPEG2000. At a similar or lower bitrate, the proposed algorithm is about 2 dB higher than the PSNR value of JPEG2000.https://www.mdpi.com/2072-4292/15/8/2211lossy compressionHRRSIlearned image compressionfused hyperpriorrate-distortion performance |
| spellingShingle | Chuan Fu Bo Du Remote Sensing Image Compression Based on the Multiple Prior Information Remote Sensing lossy compression HRRSI learned image compression fused hyperprior rate-distortion performance |
| title | Remote Sensing Image Compression Based on the Multiple Prior Information |
| title_full | Remote Sensing Image Compression Based on the Multiple Prior Information |
| title_fullStr | Remote Sensing Image Compression Based on the Multiple Prior Information |
| title_full_unstemmed | Remote Sensing Image Compression Based on the Multiple Prior Information |
| title_short | Remote Sensing Image Compression Based on the Multiple Prior Information |
| title_sort | remote sensing image compression based on the multiple prior information |
| topic | lossy compression HRRSI learned image compression fused hyperprior rate-distortion performance |
| url | https://www.mdpi.com/2072-4292/15/8/2211 |
| work_keys_str_mv | AT chuanfu remotesensingimagecompressionbasedonthemultiplepriorinformation AT bodu remotesensingimagecompressionbasedonthemultiplepriorinformation |