Conference
Evaluating INR-based and traditional image compression methods: A comparative analysis of rate-distortion performance and dataset adaptability.
| Τίτλος: | Evaluating INR-based and traditional image compression methods: A comparative analysis of rate-distortion performance and dataset adaptability. |
|---|---|
| Συγγραφείς: | Kumar, G. Prasanna, Bachanna, Prashant, Dudi, Bhanu Prakash, Khan, M. Khaleel Ullah, Kumar, M. Raman |
| Πηγή: | AIP Conference Proceedings; 2026, Vol. 3418 Issue 1, p1-7, 7p |
| Θεματικοί όροι: | Image compression, Rate distortion theory, JPEG (Image coding standard) |
| Περίληψη: | This study evaluates the performance of four image compression methods, JPEG2000, BPG, Meta-Learned Compression, and Basic Compression across two datasets, CelebA and Kodak, using Peak Signal-to-Noise Ratio (PSNR) as the primary metric. The analysis reveals that traditional methods, JPEG2000 and BPG, consistently outperform learned approaches in both rate-distortion performance and convergence speed. JPEG2000 achieves the highest PSNR values across diverse bitrates, demonstrating exceptional reliability and efficiency in preserving image quality, especially for complex datasets like Kodak. BPG excels at low bitrates, offering better visual quality under aggressive compression, though its advantage diminishes at higher bitrates. Meta-Learned Compression, a neural network-based approach, shows significant improvements over the Basic method, underscoring the potential of meta-learning techniques in improving rate-distortion efficiency. However, it remains less effective than traditional methods, with slower convergence speeds and lower PSNR values. The Basic method, serving as a baseline for INR-based techniques, exhibits the lowest performance, highlighting the need for further optimization. Dataset characteristics significantly impact performance, with CelebA's smooth textures enabling higher PSNR values, while Kodak's complex details pose greater challenges. These findings emphasize the strengths of traditional methods and highlight the promise of learned approaches as a viable future alternative, with continued innovation required to close the performance gap. [ABSTRACT FROM AUTHOR] |
| Copyright of AIP Conference Proceedings is the property of American Institute of Physics and its content may not be copied or emailed to multiple sites without the copyright holder's express written permission. Additionally, content may not be used with any artificial intelligence tools or machine learning technologies. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.) | |
| Βάση Δεδομένων: | Complementary Index |
| FullText | Text: Availability: 0 |
|---|---|
| Header | DbId: edb DbLabel: Complementary Index An: 195374546 RelevancyScore: 1076 AccessLevel: 6 PubType: Conference PubTypeId: conference PreciseRelevancyScore: 1075.76586914063 |
| IllustrationInfo | |
| Items | – Name: Title Label: Title Group: Ti Data: Evaluating INR-based and traditional image compression methods: A comparative analysis of rate-distortion performance and dataset adaptability. – Name: Author Label: Authors Group: Au Data: <searchLink fieldCode="AR" term="%22Kumar%2C+G%2E+Prasanna%22">Kumar, G. Prasanna</searchLink><br /><searchLink fieldCode="AR" term="%22Bachanna%2C+Prashant%22">Bachanna, Prashant</searchLink><br /><searchLink fieldCode="AR" term="%22Dudi%2C+Bhanu+Prakash%22">Dudi, Bhanu Prakash</searchLink><br /><searchLink fieldCode="AR" term="%22Khan%2C+M%2E+Khaleel+Ullah%22">Khan, M. Khaleel Ullah</searchLink><br /><searchLink fieldCode="AR" term="%22Kumar%2C+M%2E+Raman%22">Kumar, M. Raman</searchLink> – Name: TitleSource Label: Source Group: Src Data: AIP Conference Proceedings; 2026, Vol. 3418 Issue 1, p1-7, 7p – Name: Subject Label: Subject Terms Group: Su Data: <searchLink fieldCode="DE" term="%22Image+compression%22">Image compression</searchLink><br /><searchLink fieldCode="DE" term="%22Rate+distortion+theory%22">Rate distortion theory</searchLink><br /><searchLink fieldCode="DE" term="%22JPEG+%28Image+coding+standard%29%22">JPEG (Image coding standard)</searchLink> – Name: Abstract Label: Abstract Group: Ab Data: This study evaluates the performance of four image compression methods, JPEG2000, BPG, Meta-Learned Compression, and Basic Compression across two datasets, CelebA and Kodak, using Peak Signal-to-Noise Ratio (PSNR) as the primary metric. The analysis reveals that traditional methods, JPEG2000 and BPG, consistently outperform learned approaches in both rate-distortion performance and convergence speed. JPEG2000 achieves the highest PSNR values across diverse bitrates, demonstrating exceptional reliability and efficiency in preserving image quality, especially for complex datasets like Kodak. BPG excels at low bitrates, offering better visual quality under aggressive compression, though its advantage diminishes at higher bitrates. Meta-Learned Compression, a neural network-based approach, shows significant improvements over the Basic method, underscoring the potential of meta-learning techniques in improving rate-distortion efficiency. However, it remains less effective than traditional methods, with slower convergence speeds and lower PSNR values. The Basic method, serving as a baseline for INR-based techniques, exhibits the lowest performance, highlighting the need for further optimization. Dataset characteristics significantly impact performance, with CelebA's smooth textures enabling higher PSNR values, while Kodak's complex details pose greater challenges. These findings emphasize the strengths of traditional methods and highlight the promise of learned approaches as a viable future alternative, with continued innovation required to close the performance gap. [ABSTRACT FROM AUTHOR] – Name: Abstract Label: Group: Ab Data: <i>Copyright of AIP Conference Proceedings is the property of American Institute of Physics and its content may not be copied or emailed to multiple sites without the copyright holder's express written permission. Additionally, content may not be used with any artificial intelligence tools or machine learning technologies. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract.</i> (Copyright applies to all Abstracts.) |
| PLink | https://search.ebscohost.com/login.aspx?direct=true&site=eds-live&db=edb&AN=195374546 |
| RecordInfo | BibRecord: BibEntity: Identifiers: – Type: doi Value: 10.1063/5.0341871 Languages: – Code: eng Text: English PhysicalDescription: Pagination: PageCount: 7 StartPage: 1 Subjects: – SubjectFull: Image compression Type: general – SubjectFull: Rate distortion theory Type: general – SubjectFull: JPEG (Image coding standard) Type: general Titles: – TitleFull: Evaluating INR-based and traditional image compression methods: A comparative analysis of rate-distortion performance and dataset adaptability. Type: main BibRelationships: HasContributorRelationships: – PersonEntity: Name: NameFull: Kumar, G. Prasanna – PersonEntity: Name: NameFull: Bachanna, Prashant – PersonEntity: Name: NameFull: Dudi, Bhanu Prakash – PersonEntity: Name: NameFull: Khan, M. Khaleel Ullah – PersonEntity: Name: NameFull: Kumar, M. Raman IsPartOfRelationships: – BibEntity: Dates: – D: 18 M: 06 Text: 2026 Type: published Y: 2026 Identifiers: – Type: issn-print Value: 0094243X Numbering: – Type: volume Value: 3418 – Type: issue Value: 1 Titles: – TitleFull: AIP Conference Proceedings Type: main |
| ResultId | 1 |