A Multi-Scale Generalization of the HoG and HMAX Image Descriptors for Object Detection
Recently, several powerful image features have been proposed whichcan be described as spatial histograms of oriented energy. Forinstance, the HoG, HMAX C1, SIFT, and Shape Context feature allrepresent an input image using with a discrete set of bins whichaccumulate evidence for oriented structures o...
| Main Author: | |
|---|---|
| Other Authors: | |
| Published: |
2008
|
| Subjects: | |
| Online Access: | http://hdl.handle.net/1721.1/41093 |
| _version_ | 1811092850580914176 |
|---|---|
| author | Bileschi, Stanley M |
| author2 | Tomaso Poggio |
| author_facet | Tomaso Poggio Bileschi, Stanley M |
| author_sort | Bileschi, Stanley M |
| collection | MIT |
| description | Recently, several powerful image features have been proposed whichcan be described as spatial histograms of oriented energy. Forinstance, the HoG, HMAX C1, SIFT, and Shape Context feature allrepresent an input image using with a discrete set of bins whichaccumulate evidence for oriented structures over a spatial regionand a range of orientations. In this work, we generalize thesetechniques to allow for a foveated input image, rather than arectilinear raster. It will be shown that improved object detectionaccuracy can be achieved via inputting a spectrum of imagemeasurements, from sharp, fine-scale image sampling within a smallspatial region within the target to coarse-scale sampling of a widefield of view around the target. Several alternative featuregeneration algorithms are proposed and tested which suitably makeuse of foveated image inputs. In the experiments we show thatfeatures generated from the foveated input format produce detectorsof greater accuracy, as measured for four object types from commonlyavailable data-sets. Finally, a flexible algorithm for generatingfeatures is described and tested which is independent of inputtopology and uses ICA to learn appropriate filters. |
| first_indexed | 2024-09-23T15:30:21Z |
| id | mit-1721.1/41093 |
| institution | Massachusetts Institute of Technology |
| last_indexed | 2024-09-23T15:30:21Z |
| publishDate | 2008 |
| record_format | dspace |
| spelling | mit-1721.1/410932019-04-10T23:12:16Z A Multi-Scale Generalization of the HoG and HMAX Image Descriptors for Object Detection Bileschi, Stanley M Tomaso Poggio Center for Biological and Computational Learning (CBCL) Object Detection, ICA, Multi-Scale, Image Features Recently, several powerful image features have been proposed whichcan be described as spatial histograms of oriented energy. Forinstance, the HoG, HMAX C1, SIFT, and Shape Context feature allrepresent an input image using with a discrete set of bins whichaccumulate evidence for oriented structures over a spatial regionand a range of orientations. In this work, we generalize thesetechniques to allow for a foveated input image, rather than arectilinear raster. It will be shown that improved object detectionaccuracy can be achieved via inputting a spectrum of imagemeasurements, from sharp, fine-scale image sampling within a smallspatial region within the target to coarse-scale sampling of a widefield of view around the target. Several alternative featuregeneration algorithms are proposed and tested which suitably makeuse of foveated image inputs. In the experiments we show thatfeatures generated from the foveated input format produce detectorsof greater accuracy, as measured for four object types from commonlyavailable data-sets. Finally, a flexible algorithm for generatingfeatures is described and tested which is independent of inputtopology and uses ICA to learn appropriate filters. 2008-04-09T20:15:10Z 2008-04-09T20:15:10Z 2008-04-09 MIT-CSAIL-TR-2008-019 CBCL-271 http://hdl.handle.net/1721.1/41093 Massachusetts Institute of Technology Computer Science and Artificial Intelligence Laboratory 8 p. application/pdf application/postscript |
| spellingShingle | Object Detection, ICA, Multi-Scale, Image Features Bileschi, Stanley M A Multi-Scale Generalization of the HoG and HMAX Image Descriptors for Object Detection |
| title | A Multi-Scale Generalization of the HoG and HMAX Image Descriptors for Object Detection |
| title_full | A Multi-Scale Generalization of the HoG and HMAX Image Descriptors for Object Detection |
| title_fullStr | A Multi-Scale Generalization of the HoG and HMAX Image Descriptors for Object Detection |
| title_full_unstemmed | A Multi-Scale Generalization of the HoG and HMAX Image Descriptors for Object Detection |
| title_short | A Multi-Scale Generalization of the HoG and HMAX Image Descriptors for Object Detection |
| title_sort | multi scale generalization of the hog and hmax image descriptors for object detection |
| topic | Object Detection, ICA, Multi-Scale, Image Features |
| url | http://hdl.handle.net/1721.1/41093 |
| work_keys_str_mv | AT bileschistanleym amultiscalegeneralizationofthehogandhmaximagedescriptorsforobjectdetection AT bileschistanleym multiscalegeneralizationofthehogandhmaximagedescriptorsforobjectdetection |