Object detection using binary coded images and multi-stage cascade classifiers
AssigneeIntel Corporation
InventorsJianguo Li, Ke Chen, Yi-Jen Chiu, Chen Wang, Ya-Ti Peng, Bin Wang
Life of this patent
Estimating the term…
Abstract
Techniques related to object detection using directional filtering are discussed. Such techniques may include determining directional weighted averages for pixels of an input image, generating a feature representation of the input image based on the directional weighted averages, and performing object detection by applying a multi-stage cascade classifier to the feature representation.
Claims19
A computer-implemented method for object detection comprising:
- determining directional weighted averages for individual pixels of an input image based at least in part on pixel values of a plurality of neighbor pixels of each of the individual pixels and weights associated with locations of each of the neighbors pixels relative to the individual pixels;
- generating a binary coded image for the input image by assigning, for each of the individual pixels, the neighbor pixels thereof one of a first value or second value based on a comparison of each pixel value of the neighbor pixels to the directional weighted average of the individual pixel and determining a binary coded representation for each of the individual pixels by arranging the first or second values of the neighbor pixels thereof into an order;
- performing object detection for the input image by applying a multi-stage cascade classifier to the binary coded image.
A system for providing object detection on a computing device, comprising:
- a display device configured to present image data;
- and
- a graphics processing unit coupled to the display device, the graphics processing unit to:
- determine directional weighted averages for individual pixels of an input image based at least in part on pixel values of a plurality of neighbor pixels of each of the individual pixels and weights associated with locations of each of the neighbors pixels relative to the individual pixels;
- generate a binary coded image for the input image by assignment, for each of the individual pixels, the neighbor pixels thereof one of a first value or second value based on a comparison of each pixel value of the neighbor pixels to the directional weighted average of the individual pixel and determination of a binary coded representation for each of the individual pixels by arrangement of the first or second values of the neighbor pixels thereof into an order;
- and
- perform object detection for the input image by applying a multi-stage cascade classifier to the binary coded image.
At least one machine readable medium comprising a plurality of instructions that in response to being executed on a computing device, cause the computing device to perform object detection by:
- determining directional weighted averages for individual pixels of an input image based at least in part on pixel values of a plurality of neighbor pixels of each of the individual pixels and weights associated with locations of each of the neighbors pixels relative to the individual pixels;
- generating a binary coded image for the input image by assigning, for each of the individual pixels, the neighbor pixels thereof one of a first value or second value based on a comparison of each pixel value of the neighbor pixels to the directional weighted average of the individual pixel and determining a binary coded representation for each of the individual pixels by arranging the first or second values of the neighbor pixels thereof into an order;
- performing object detection for the input image by applying a multi-stage cascade classifier to the binary coded image.
Dependencies are inferred from the claim text.