Systems and methods for providing an image classifier

1. A computing device for image classification comprising:

a memory including:

a non-rectangular foreground silhouette generated based on a set of training images, and

a non-rectangular bitmap that identifies a subset of cells of a rectangular detection window, the rectangular detection window having a cell width and cell height;

a hardware accelerator to:

identify a portion of an image to be classified based on the non-rectangular foreground silhouette,

identify a cell within the portion of the image to be classified that match the subset of the cells of the non-rectangular bitmap, the portion having the cell width and the cell height,

generate and store a histogram of orientations of pixels within the cell that matches the subset of cells, and

determine a cell block having a plurality of contiguous identified cells from the image; and

a processor to:

generate, for the cell block, a descriptor including at least a portion of the histogram, and

evaluate the descriptor against the set of training images to selectively identify the portion of the image as including a class of object associated with at least one of the training images.

2. The computing device of claim 1 , wherein the hardware accelerator is to:

re-scale the image to a second cell width and cell height; and

apply the rectangular detection window to a portion of the re-scaled image having the cell width and cell height.

3. The computing device of claim 1 , wherein at least one of the hardware accelerator or the processor is to:

average the training images; and

store the averaged training images as the non-rectangular foreground silhouette.

4. The computing device of claim 1 , wherein at least one of the hardware accelerator or the processor is to:

determine intensity values of pixels for the training images;

add the intensity values of the pixels for the training images to generate resulting values;

divide the resulting values by a number of the training images; and

store the divided resulting values as the non-rectangular foreground silhouette.

5. The computing device of claim 1 , wherein the hardware accelerator is to:

calculate gradients, the gradients having magnitudes and corresponding orientations;

associate the orientations to respective ones of a plurality of bins; and

add the magnitudes to the respective ones of the bins associated with corresponding ones of the orientations.

6. The computing device of claim 1 , wherein the processor is to determine whether the class of object is one of:

a person;

a face;

a non-rectangular geometric shape; or

a non-geometric shape.

7. The computing device of claim 1 , wherein determining the cell block includes determining a plurality of overlapping cell blocks having a plurality of contiguous identified cells.

8. A non-transitory computer readable medium comprising instructions that, when executed, cause a computing device to at least:

store a non-rectangular bitmap that identifies a subset of cells of a rectangular detection window, the rectangular detection window having a cell width and cell height;

identify a portion of an image to be classified based on a non-rectangular foreground silhouette, the non-rectangular foreground silhouette generated based on a set of training images;

identify a cell within the portion of the image to be classified that match the subset of the cells of the non-rectangular bitmap, the portion having the cell width and the cell height;

generate and store a histogram of orientations of pixels within the cell that matches the subset of cells;

determine a cell block from the image;

generate, for the cell block, a descriptor including at least a portion of the histogram; and

evaluate the descriptor based on the set of training images to identify the portion of the image as including a class of object associated with at least one of the training images.

9. The non-transitory computer readable medium of claim 8 , wherein the instructions, when executed, cause the computing device to:

re-scale the image to a second cell width and cell height; and

apply the rectangular detection window to a portion of the re-scaled image having the cell width and cell height.

10. The non-transitory computer readable medium of claim 8 , wherein the instructions, when executed, cause the computing device to:

average the training images; and

store the averaged training images as the non-rectangular foreground silhouette.

11. The non-transitory computer readable medium of claim 8 , wherein the instructions, when executed, cause the computing device to:

determine intensity values of pixels for the training images;

add the intensity values of the pixels for the training images to generate resulting values;

divide the resulting values by a number of the training images; and

store the divided resulting values as the non-rectangular foreground silhouette.

12. The non-transitory computer readable medium of claim 8 , wherein the instructions, when executed, cause the computing device to:

calculate gradients, the gradients having magnitudes and corresponding orientations;

associate the orientations to respective ones of a plurality of bins; and

add the magnitudes to the respective ones of the bins associated with corresponding ones of the orientations.

13. The non-transitory computer readable medium of claim 8 , wherein the instructions, when executed, cause the computing device to determine whether the class of object is one of:

a person;

a face;

a non-rectangular geometric shape; or

a non-geometric shape.

14. The non-transitory computer readable medium of claim 8 , wherein the instructions, when executed, cause the computing device to determine the cell block by determining a plurality of overlapping cell blocks having a plurality of contiguous identified cells.

15. A method comprising:

storing a non-rectangular foreground silhouette generated based on a set of training images;

storing a non-rectangular bitmap that identifies a subset of cells of a rectangular detection window, the rectangular detection window having a cell width and cell height;

identifying a portion of an image to be classified based on the non-rectangular foreground silhouette;

identifying a cell within the portion of the image to be classified that match the subset of the cells of the non-rectangular bitmap, the portion having the cell width and the cell height;

generating and store a histogram of orientations of pixels within the cell that matches the subset of cells;

determining a cell block having a plurality of contiguous identified cells from the image;

generating, for the cell block, a descriptor including at least a portion of the histogram; and

evaluating the descriptor against the set of training images to selectively identify the portion of the image as including a class of object associated with at least one of the training images.

16. The method of claim 15 , further including:

re-scaling the image to a second cell width and cell height; and

applying the rectangular detection window to a portion of the re-scaled image having the cell width and cell height.

17. The method of claim 15 , further including:

averaging the training images; and

storing the averaged training images as the non-rectangular foreground silhouette.

18. The method of claim 15 , further including:

determining intensity values of pixels for the training images;

adding the intensity values of the pixels for the training images to generate resulting values;

dividing the resulting values by a number of the training images; and

storing the divided resulting values as the non-rectangular foreground silhouette.

19. The method of claim 15 , further including:

calculating gradients, the gradients having magnitudes and corresponding orientations;

associating the orientations to respective ones of a plurality of bins; and

adding the magnitudes to the respective ones of the bins associated with corresponding ones of the orientations.

20. The method of claim 15 , further including determining whether the class of object is one of:

a person;

a face;

a non-rectangular geometric shape; or

a non-geometric shape.

21. The method of claim 15 , wherein determining the cell block includes determining a plurality of overlapping cell blocks having a plurality of contiguous identified cells.

22. A computing device for image classification comprising:

means for storing:

a non-rectangular foreground silhouette generated based on a set of training images, and

a non-rectangular bitmap that identifies a subset of cells of a rectangular detection window, the rectangular detection window having a cell width and cell height;

means for determining a cell block to:

identify a portion of an image to be classified based on the non-rectangular foreground silhouette,

identify a cell within the portion of the image to be classified that match the subset of the cells of the non-rectangular bitmap, the portion having the cell width and the cell height,

generate and store a histogram of orientations of pixels within the cell that matches the subset of cells, and

determine a cell block having a plurality of contiguous identified cells from the image; and

means for identifying to:

generate, for the cell block, a descriptor including at least a portion of the histogram, and

evaluate the descriptor against the set of training images to selectively identify the portion of the image as including a class of object associated with at least one of the training images.

23. The computing device of claim 22 , wherein the means for determining the cell block is to:

re-scale the image to a second cell width and cell height; and

apply the rectangular detection window to a portion of the re-scaled image having the cell width and cell height.

24. The computing device of claim 22 , wherein the means for determining the cell block is to:

average the training images; and

store the averaged training images as the non-rectangular foreground silhouette.

25. The computing device of claim 22 , wherein the means for determining the cell block is to:

determine intensity values of pixels for the training images;

add the intensity values of the pixels for the training images to generate resulting values;

divide the resulting values by a number of the training images; and

store the divided resulting values as the non-rectangular foreground silhouette.

26. The computing device of claim 22 , further including means for determining whether the class of object is one of:

a person;

a face;

a non-rectangular geometric shape; or

a non-geometric shape.

27. The computing device of claim 22 , wherein the means for determining the cell block is to determine a plurality of overlapping cell blocks having a plurality of contiguous identified cells.