Perceptually-based Comparison of Image Similarity Metrics
The image comparison operation ??sessing how well one image matches another ??rms a critical component of many image analysis systems and models of human visual processing. Two norms used commonly for this purpose are L1 and L2, which are specific instances of the Minkowski metric. However, there is...
| Main Authors: | , |
|---|---|
| Language: | en_US |
| Published: |
2004
|
| Subjects: | |
| Online Access: | http://hdl.handle.net/1721.1/7235 |
| _version_ | 1826203747744743424 |
|---|---|
| author | Russell, Richard Sinha, Pawan |
| author_facet | Russell, Richard Sinha, Pawan |
| author_sort | Russell, Richard |
| collection | MIT |
| description | The image comparison operation ??sessing how well one image matches another ??rms a critical component of many image analysis systems and models of human visual processing. Two norms used commonly for this purpose are L1 and L2, which are specific instances of the Minkowski metric. However, there is often not a principled reason for selecting one norm over the other. One way to address this problem is by examining whether one metric better captures the perceptual notion of image similarity than the other. With this goal, we examined perceptual preferences for images retrieved on the basis of the L1 versus the L2 norm. These images were either small fragments without recognizable content, or larger patterns with recognizable content created via vector quantization. In both conditions the subjects showed a consistent preference for images matched using the L1 metric. These results suggest that, in the domain of natural images of the kind we have used, the L1 metric may better capture human notions of image similarity. |
| first_indexed | 2024-09-23T12:42:25Z |
| id | mit-1721.1/7235 |
| institution | Massachusetts Institute of Technology |
| language | en_US |
| last_indexed | 2024-09-23T12:42:25Z |
| publishDate | 2004 |
| record_format | dspace |
| spelling | mit-1721.1/72352019-04-12T08:34:08Z Perceptually-based Comparison of Image Similarity Metrics Russell, Richard Sinha, Pawan AI Image matching vector quantization Minkowski metric The image comparison operation ??sessing how well one image matches another ??rms a critical component of many image analysis systems and models of human visual processing. Two norms used commonly for this purpose are L1 and L2, which are specific instances of the Minkowski metric. However, there is often not a principled reason for selecting one norm over the other. One way to address this problem is by examining whether one metric better captures the perceptual notion of image similarity than the other. With this goal, we examined perceptual preferences for images retrieved on the basis of the L1 versus the L2 norm. These images were either small fragments without recognizable content, or larger patterns with recognizable content created via vector quantization. In both conditions the subjects showed a consistent preference for images matched using the L1 metric. These results suggest that, in the domain of natural images of the kind we have used, the L1 metric may better capture human notions of image similarity. 2004-10-20T21:03:39Z 2004-10-20T21:03:39Z 2001-07-01 AIM-2001-014 CBCL-201 http://hdl.handle.net/1721.1/7235 en_US AIM-2001-014 CBCL-201 13 p. 9714300 bytes 2612761 bytes application/postscript application/pdf application/postscript application/pdf |
| spellingShingle | AI Image matching vector quantization Minkowski metric Russell, Richard Sinha, Pawan Perceptually-based Comparison of Image Similarity Metrics |
| title | Perceptually-based Comparison of Image Similarity Metrics |
| title_full | Perceptually-based Comparison of Image Similarity Metrics |
| title_fullStr | Perceptually-based Comparison of Image Similarity Metrics |
| title_full_unstemmed | Perceptually-based Comparison of Image Similarity Metrics |
| title_short | Perceptually-based Comparison of Image Similarity Metrics |
| title_sort | perceptually based comparison of image similarity metrics |
| topic | AI Image matching vector quantization Minkowski metric |
| url | http://hdl.handle.net/1721.1/7235 |
| work_keys_str_mv | AT russellrichard perceptuallybasedcomparisonofimagesimilaritymetrics AT sinhapawan perceptuallybasedcomparisonofimagesimilaritymetrics |