Learning Object-Independent Modes of Variation with Feature Flow Fields
We present a unifying framework in which "object-independent" modes of variation are learned from continuous-time data such as video sequences. These modes of variation can be used as "generators" to produce a manifold of images of a new object from a single example of that obje...
| Main Authors: | , , |
|---|---|
| Language: | en_US |
| Published: |
2004
|
| Subjects: | |
| Online Access: | http://hdl.handle.net/1721.1/6659 |
| _version_ | 1826216610750267392 |
|---|---|
| author | Miller, Erik G. Tieu, Kinh Stauffer, Chris P. |
| author_facet | Miller, Erik G. Tieu, Kinh Stauffer, Chris P. |
| author_sort | Miller, Erik G. |
| collection | MIT |
| description | We present a unifying framework in which "object-independent" modes of variation are learned from continuous-time data such as video sequences. These modes of variation can be used as "generators" to produce a manifold of images of a new object from a single example of that object. We develop the framework in the context of a well-known example: analyzing the modes of spatial deformations of a scene under camera movement. Our method learns a close approximation to the standard affine deformations that are expected from the geometry of the situation, and does so in a completely unsupervised (i.e. ignorant of the geometry of the situation) fashion. We stress that it is learning a "parameterization", not just the parameter values, of the data. We then demonstrate how we have used the same framework to derive a novel data-driven model of joint color change in images due to common lighting variations. The model is superior to previous models of color change in describing non-linear color changes due to lighting. |
| first_indexed | 2024-09-23T16:49:53Z |
| id | mit-1721.1/6659 |
| institution | Massachusetts Institute of Technology |
| language | en_US |
| last_indexed | 2024-09-23T16:49:53Z |
| publishDate | 2004 |
| record_format | dspace |
| spelling | mit-1721.1/66592019-04-11T02:52:43Z Learning Object-Independent Modes of Variation with Feature Flow Fields Miller, Erik G. Tieu, Kinh Stauffer, Chris P. AI Invariance Optical Flow Color Constancy Object Recognition image manifold We present a unifying framework in which "object-independent" modes of variation are learned from continuous-time data such as video sequences. These modes of variation can be used as "generators" to produce a manifold of images of a new object from a single example of that object. We develop the framework in the context of a well-known example: analyzing the modes of spatial deformations of a scene under camera movement. Our method learns a close approximation to the standard affine deformations that are expected from the geometry of the situation, and does so in a completely unsupervised (i.e. ignorant of the geometry of the situation) fashion. We stress that it is learning a "parameterization", not just the parameter values, of the data. We then demonstrate how we have used the same framework to derive a novel data-driven model of joint color change in images due to common lighting variations. The model is superior to previous models of color change in describing non-linear color changes due to lighting. 2004-10-08T20:36:37Z 2004-10-08T20:36:37Z 2001-09-01 AIM-2001-021 http://hdl.handle.net/1721.1/6659 en_US AIM-2001-021 9 p. 8233900 bytes 814636 bytes application/postscript application/pdf application/postscript application/pdf |
| spellingShingle | AI Invariance Optical Flow Color Constancy Object Recognition image manifold Miller, Erik G. Tieu, Kinh Stauffer, Chris P. Learning Object-Independent Modes of Variation with Feature Flow Fields |
| title | Learning Object-Independent Modes of Variation with Feature Flow Fields |
| title_full | Learning Object-Independent Modes of Variation with Feature Flow Fields |
| title_fullStr | Learning Object-Independent Modes of Variation with Feature Flow Fields |
| title_full_unstemmed | Learning Object-Independent Modes of Variation with Feature Flow Fields |
| title_short | Learning Object-Independent Modes of Variation with Feature Flow Fields |
| title_sort | learning object independent modes of variation with feature flow fields |
| topic | AI Invariance Optical Flow Color Constancy Object Recognition image manifold |
| url | http://hdl.handle.net/1721.1/6659 |
| work_keys_str_mv | AT millererikg learningobjectindependentmodesofvariationwithfeatureflowfields AT tieukinh learningobjectindependentmodesofvariationwithfeatureflowfields AT staufferchrisp learningobjectindependentmodesofvariationwithfeatureflowfields |