Random Lens Imaging
We call a random lens one for which the function relating the input light ray to the output sensor location is pseudo-random. Imaging systems with random lensescan expand the space of possible camera designs, allowing new trade-offs in optical design and potentially adding new imaging capabilities....
| Main Authors: | , , |
|---|---|
| Other Authors: | |
| Language: | en_US |
| Published: |
2006
|
| Online Access: | http://hdl.handle.net/1721.1/33962 |
| _version_ | 1811084041845211136 |
|---|---|
| author | Fergus, Rob Torralba, Antonio Freeman, William T. |
| author2 | William Freeman |
| author_facet | William Freeman Fergus, Rob Torralba, Antonio Freeman, William T. |
| author_sort | Fergus, Rob |
| collection | MIT |
| description | We call a random lens one for which the function relating the input light ray to the output sensor location is pseudo-random. Imaging systems with random lensescan expand the space of possible camera designs, allowing new trade-offs in optical design and potentially adding new imaging capabilities. Machine learningmethods are critical for both camera calibration and image reconstruction from the sensor data. We develop the theory and compare two different methods for calibration and reconstruction: an MAP approach, and basis pursuit from compressive sensing. We show proof-of-concept experimental results from a random lens made from a multi-faceted mirror, showing successful calibration and image reconstruction. We illustrate the potential for super-resolution and 3D imaging. |
| first_indexed | 2024-09-23T12:43:43Z |
| id | mit-1721.1/33962 |
| institution | Massachusetts Institute of Technology |
| language | en_US |
| last_indexed | 2024-09-23T12:43:43Z |
| publishDate | 2006 |
| record_format | dspace |
| spelling | mit-1721.1/339622019-04-12T08:37:57Z Random Lens Imaging Fergus, Rob Torralba, Antonio Freeman, William T. William Freeman Vision We call a random lens one for which the function relating the input light ray to the output sensor location is pseudo-random. Imaging systems with random lensescan expand the space of possible camera designs, allowing new trade-offs in optical design and potentially adding new imaging capabilities. Machine learningmethods are critical for both camera calibration and image reconstruction from the sensor data. We develop the theory and compare two different methods for calibration and reconstruction: an MAP approach, and basis pursuit from compressive sensing. We show proof-of-concept experimental results from a random lens made from a multi-faceted mirror, showing successful calibration and image reconstruction. We illustrate the potential for super-resolution and 3D imaging. 2006-09-07T18:46:57Z 2006-09-07T18:46:57Z 2006-09-02 MIT-CSAIL-TR-2006-058 http://hdl.handle.net/1721.1/33962 en_US Massachusetts Institute of Technology Computer Science and Artificial Intelligence Laboratory 9 p. 52671890 bytes 1133927 bytes application/postscript application/pdf application/postscript application/pdf |
| spellingShingle | Fergus, Rob Torralba, Antonio Freeman, William T. Random Lens Imaging |
| title | Random Lens Imaging |
| title_full | Random Lens Imaging |
| title_fullStr | Random Lens Imaging |
| title_full_unstemmed | Random Lens Imaging |
| title_short | Random Lens Imaging |
| title_sort | random lens imaging |
| url | http://hdl.handle.net/1721.1/33962 |
| work_keys_str_mv | AT fergusrob randomlensimaging AT torralbaantonio randomlensimaging AT freemanwilliamt randomlensimaging |