Overcoming registration uncertainty in image super-resolution: maximize or marginalize?
In multiple-image super-resolution, a high-resolution image is estimated from a number of lower-resolution images. This usually involves computing the parameters of a generative imaging model (such as geometric and photometric registration, and blur) and obtaining a MAP estimate by minimizing a cost...
| Main Authors: | , , , |
|---|---|
| Format: | Journal article |
| Language: | English |
| Published: |
Springer
2007
|
| _version_ | 1824459023973351424 |
|---|---|
| author | Pickup, L Capel, D Roberts, S Zisserman, A |
| author_facet | Pickup, L Capel, D Roberts, S Zisserman, A |
| author_sort | Pickup, L |
| collection | OXFORD |
| description | In multiple-image super-resolution, a high-resolution image is estimated from a number of lower-resolution images. This usually involves computing the parameters of a generative imaging model (such as geometric and photometric registration, and blur) and obtaining a MAP estimate by minimizing a cost function including an appropriate prior. Two alternative approaches are examined. First, both registrations and the super-resolution image are found simultaneously using a joint MAP optimization. Second, we perform Bayesian integration over the unknown image registration parameters, deriving a cost function whose only variables of interest are the pixel values of the super-resolution image. We also introduce a scheme to learn the parameters of the image prior as part of the super-resolution algorithm. We show examples on a number of real sequences including multiple stills, digital video, and DVDs of movies. |
| first_indexed | 2024-03-07T02:48:00Z |
| format | Journal article |
| id | oxford-uuid:acadae6b-a250-4064-9b0c-4ee290dc4404 |
| institution | University of Oxford |
| language | English |
| last_indexed | 2025-02-19T04:35:12Z |
| publishDate | 2007 |
| publisher | Springer |
| record_format | dspace |
| spelling | oxford-uuid:acadae6b-a250-4064-9b0c-4ee290dc44042025-01-23T12:55:09ZOvercoming registration uncertainty in image super-resolution: maximize or marginalize?Journal articlehttp://purl.org/coar/resource_type/c_dcae04bcuuid:acadae6b-a250-4064-9b0c-4ee290dc4404EnglishSymplectic Elements at OxfordSpringer2007Pickup, LCapel, DRoberts, SZisserman, AIn multiple-image super-resolution, a high-resolution image is estimated from a number of lower-resolution images. This usually involves computing the parameters of a generative imaging model (such as geometric and photometric registration, and blur) and obtaining a MAP estimate by minimizing a cost function including an appropriate prior. Two alternative approaches are examined. First, both registrations and the super-resolution image are found simultaneously using a joint MAP optimization. Second, we perform Bayesian integration over the unknown image registration parameters, deriving a cost function whose only variables of interest are the pixel values of the super-resolution image. We also introduce a scheme to learn the parameters of the image prior as part of the super-resolution algorithm. We show examples on a number of real sequences including multiple stills, digital video, and DVDs of movies. |
| spellingShingle | Pickup, L Capel, D Roberts, S Zisserman, A Overcoming registration uncertainty in image super-resolution: maximize or marginalize? |
| title | Overcoming registration uncertainty in image super-resolution: maximize or marginalize? |
| title_full | Overcoming registration uncertainty in image super-resolution: maximize or marginalize? |
| title_fullStr | Overcoming registration uncertainty in image super-resolution: maximize or marginalize? |
| title_full_unstemmed | Overcoming registration uncertainty in image super-resolution: maximize or marginalize? |
| title_short | Overcoming registration uncertainty in image super-resolution: maximize or marginalize? |
| title_sort | overcoming registration uncertainty in image super resolution maximize or marginalize |
| work_keys_str_mv | AT pickupl overcomingregistrationuncertaintyinimagesuperresolutionmaximizeormarginalize AT capeld overcomingregistrationuncertaintyinimagesuperresolutionmaximizeormarginalize AT robertss overcomingregistrationuncertaintyinimagesuperresolutionmaximizeormarginalize AT zissermana overcomingregistrationuncertaintyinimagesuperresolutionmaximizeormarginalize |