Three dimensional (3D) object detection

1. A circuit configured to:

receive a video image;

compute an optical flow image, based on the video image;

compute a magnitude of optical flow image for the video image based on the optical flow image; and

compute a histogram of optical flow magnitudes (HOFM) image for the video image based on the magnitude of optical flow image.

2. The circuit of claim 1 , wherein the circuit is further configured to generate a mask indicating a set of regions of interest (ROIs) of the video image, based on the HOFM.

3. The circuit of claim 2 , wherein the circuit is further configured to:

compute a histogram of normalized optical flow gradients (HOFG) image for the video image based on the HOFM image for the video image and the mask, in the set of ROIs of the video image;

classify regions of the video image as containing three dimensional (3D) objects, flat features, or no features, based on the HOFG image for the video image, to generate an object classification; and output the object classification.

4. The circuit of claim 1 , wherein computing the HOFM image includes computing the HOFM image over a first set of overlapping blocks at a first scale and over a second set of overlapping blocks at a second scale.

5. The circuit of claim 1 , wherein computing the HOFM image includes computing the HOFM image over a first set of non-overlapping blocks at a first scale and over a second set of non-overlapping blocks at a second scale.

6. The circuit of claim 1 , wherein the circuit is further configured to:

perform object detection based on the HOFM image to generate an object detection result; and

output the object detection result.

7. The circuit of claim 1 , wherein the circuit comprises a processor.

8. The circuit of claim 1 , wherein the circuit comprises a histogram generation circuit.

9. A circuit configured to:

obtain a video image;

compute an optical flow image for the video image;

compute a gradient of optical flow image based on the optical flow image; and

compute a histogram of normalized optical flow gradients (HOFG) image for the gradient of optical flow image.

10. The circuit of claim 9 , wherein the circuit is further configured to:

classify regions of the video image as containing three dimensional (3D) objects, flat features, or no object, based on the HOFG image, to generate an object classification; and

output the object classification.

11. The circuit of claim 9 , wherein the circuit is further configured to:

obtain a mask indicating regions of interest (ROIs) of the optical flow image; and

wherein computing the HOFG image is performed in the ROIs of the optical flow image.

12. The circuit of claim 9 , wherein the circuit is further configured to:

select a first set of overlapping blocks of the optical flow image at a first scale; and

select a second set of overlapping blocks of the optical flow image at a second scale; and

wherein computing the HOFG image for the video image is performed on the first set of overlapping blocks and on the second set of overlapping blocks.

13. The circuit of claim 9 , wherein the circuit is further configured to:

select a first set of non-overlapping blocks of the optical flow image at a first scale; and

select a second set of non-overlapping blocks of the optical flow image at a second scale; and

wherein computing the HOFG image for the video image is performed on the first set of non-overlapping blocks and on the second set of non-overlapping blocks.

14. The circuit of claim 9 , wherein the circuit comprises a processor.

15. The circuit of claim 9 , wherein the circuit comprises a histogram generation circuit.

16. A vehicle comprising

a camera configured to receive a video image;

a circuit coupled to the camera, the circuit configured to:

compute an optical flow image, based on the video image;

compute a magnitude of optical flow image for the video image based on the optical flow image; and

compute a histogram of optical flow magnitudes (HOFM) image for the video image based on the magnitude of optical flow image.

17. The vehicle of claim 16 , further comprising an imaging and video accelerator (IVA) circuit configured to compute the optical flow image.

18. The vehicle of claim 16 , further comprising a histogram generation circuit configured to:

compute the magnitude of optical flow image; and

compute the HOFM image.

19. The vehicle of claim 18 , wherein the histogram generation circuit comprises:

a first register block comprising an output;

a second register block comprising an output;

an M way comparator comprising:

a first input coupled to the output of the first register block;

a second input coupled to the output of the second register block; and

an output

a histogram merging circuit comprising an input coupled to the output of the M way comparator.

20. The vehicle of claim 16 , further comprising a processor configured to:

compute the magnitude of optical flow image; and

compute the HOFM image.