Face image prioritization based on face quality analysis

1. A method for face image prioritization based on face image prioritization based on face quality analysis:

detecting a facial image from a first track of facial images;

filtering the facial image with a first number of quality analysis filters that include a first number of quality analysis factors;

passing the facial image to a second number of quality analysis filters that includes a second number of quality analysis factors when the facial image passes the first number of quality analysis filters, wherein processing associated with the second number of quality analysis factors is more complex than processing associated with the first number of quality analysis factors;

determining an overall image quality of the facial image based on the first number of quality analysis factors and the second number of quality analysis factors;

determining a matching confidence between the facial image and a stored facial image with a facial recognition program based on the overall image quality; and

assigning a priority to the first track of facial images based on the matching confidence and the overall image quality of the facial image; and

determining whether to continue processing images from the first track of facial images by comparing the priority of the first track of facial images with a priority of a second track of facial images.

2. The method of claim 1 , wherein at least one of the first and second number of quality analysis factors includes at least one of face size; brightness of the facial image; dynamic range of pixels on the facial image; a sharpness metric; a blur estimate; landmark estimates from eyes, nose, and mouth; facial pose; a shape fitting residual; and an overall image quality.

3. The method of claim 2 , further comprising arranging the first number of quality analysis filters in order of increasing complexity of the quality analysis factors associated with each of the number of quality analysis filters.

4. The method of claim 1 , further comprising discarding the facial image when the facial image does not pass at least one of the number of quality analysis filters.

5. The method of claim 1 , further comprising receiving identity data from the facial recognition program in response to submitting the facial image to the facial recognition program when the facial image passes the first number of quality analysis filters.

6. The method of claim 1 , wherein filtering the facial image with the first number of quality analysis filters includes filtering the facial image with a plurality of quality analysis filters.

7. The method of claim 1 , further comprising determining whether an overall image quality of the facial image has converged.

8. A machine-readable non-transitory medium storing instructions for face image prioritization based on face quality analysis, executable by a machine to cause the machine to:

detect a facial image from a first track of facial images;

filter the facial image with a first number of quality analysis filters that include a first number of quality analysis factors;

pass the facial image to a second number of quality analysis filters that includes a second number of quality analysis factors when the facial image passes the first number of quality analysis filters, wherein processing associated with the second number of quality analysis factors is more complex than processing associated with the first number of quality analysis factors;

determine an overall image quality of the facial image based on the first number of quality analysis factors and the second number of quality analysis factors;

determine a matching confidence between the facial image and a stored facial image with a facial recognition program based on the overall image quality; and

assign a priority to a first track of facial images based on the matching confidence and the overall image quality of the facial image; and

determine whether to continue processing images from the first track of facial images by comparing the priority of the first track of facial images with a priority of a second track of facial images.

9. The medium of claim 8 , wherein the first number of quality analysis factors includes at least one of face size, brightness of the facial image, dynamic range of pixels on the facial image, and a sharpness metric.

10. The medium of claim 9 , further comprising instructions executable by the machine to arrange the first number of quality analysis filters in order of increasing complexity of the quality analysis factors associated with each of the number of quality analysis filters.

11. The medium of claim 8 , wherein the second number of quality analysis factors includes at least one or a blur estimate; landmark estimates from eyes, nose, and mouth; facial pose; a shape fitting residual; and an overall image quality.

12. The medium of claim 11 , further comprising instructions executable by the machine to arrange the second number of quality analysis filters in order of increasing complexity of the quality analysis factors associated with each of the second number of quality analysis filters.

13. The medium of claim 8 , further comprising instructions executable by the machine to process images of the second track of facial images when the priority of the second track of facial images is greater than the priority of the first track of facial images.

14. The medium of claim 8 , further comprising instructions executable by the machine to determine when overall image quality of the facial image will improve based on a model of overall image quality; and

further comprising instructions executable by the machine to delay filtering of a facial image based on the determination of when overall image quality of the facial image will improve.

15. The medium of claim 8 , further comprising instructions executable by the machine to assign a priority to the first track based on at least one of face size in the facial image and face speed in the facial image.

16. The method of claim 1 , further comprising:

storing a number of facial images that have an overall image quality exceeding a threshold image quality; and

discarding additional images that surpass the threshold image quality.

17. The method of claim 16 , further comprising varying the number of facial images that are stored based on demands for facial image processing.