Method, apparatus and computer program product for providing an orientation independent face detector

1. A method comprising:

generating multiple mosaic images from an input image, each of the multiple mosaic images having a different scale;

employing a plurality of differently oriented edge detectors to perform edge detection on the multiple mosaic images including combining edges of the multiple mosaic images having the different scales; and

performing face detection in regions corresponding to the differently oriented edge detectors based on respective feature maps produced by the differently oriented edge detectors.

2. The method of claim 1 , further comprising performing a verification of a candidate face detected by the face detection.

3. The method of claim 1 , wherein employing the plurality of differently oriented edge detectors comprises utilizing edge detectors oriented with equal amounts of angular separation between each edge detector and at least one other edge detector and oriented to cover three hundred and sixty degrees of possible in-plane rotation angles for candidate faces in respective swaths of equal size.

4. The method of claim 1 , wherein performing face detection comprises employing an elastic face template to data in the respective feature maps to detect a candidate face based on features of the candidate face matching the elastic face template.

5. The method of claim 1 , wherein performing face detection comprises employing an elastic face template having a width corresponding to a distance between eyes of a detected eye pair and a height corresponding to a distance between a line connecting the detected eye pair and a potential mouth of a candidate face.

6. An apparatus comprising a processor and a memory device including instructions, the memory device and the instructions configured to, with the processor, cause the apparatus at least to:

generate multiple mosaic images from an input image, each of the multiple mosaic images having a different scale;

employ a plurality of differently oriented edge detectors to perform edge detection on the multiple mosaic images including combining edges of the multiple mosaic images having the different scales; and

perform face detection in regions corresponding to the differently oriented edge detectors based on respective feature maps produced by the differently oriented edge detectors.

7. The apparatus of claim 6 , wherein the processor is further configured to perform a verification of a candidate face detected by the face detection.

8. The apparatus of claim 7 , wherein the processor is further configured to rotate the candidate face to a vertical orientation prior to performing the verification.

9. The apparatus of claim 7 , wherein the processor is configured to perform the verification by employing an Adaboost verification with greater than four kinds of haar-like features.

10. The apparatus of claim 7 , wherein the processor is configured to perform the verification by employing bootstrapping with a punishment coefficient applied to samples proximate to a cascade margin.

11. The apparatus of claim 6 , wherein the processor is configured to employ the plurality of differently oriented edge detectors by utilizing edge detectors oriented with equal amounts of angular separation between each edge detector and at least one other edge detector and oriented to cover three hundred and sixty degrees of possible in-plane rotation angles for candidate faces in respective swaths of equal size.

12. The apparatus of claim 6 , wherein the processor is configured to perform face detection by employing an elastic face template to data in the respective feature maps to detect a candidate face based on features of the candidate face matching the elastic face template.

13. The apparatus of claim 6 , wherein the processor is configured to perform face detection by employing an elastic face template having a width corresponding to a distance between eyes of a detected eye pair and a height corresponding to a distance between a line connecting the detected eye pair and a potential mouth of a candidate face.

14. The apparatus of claim 6 , wherein the processor is configured to generate multiple mosaic images by averaging pixel values in each of a plurality of cell windows to generate a mosaic representation of the input image comprising the plurality of cell windows and utilizing multiple levels to binarize mosaic images having different scales.

15. The apparatus of claim 6 , wherein the processor is configured to perform face detection by employing multiple thresholds on respective ones of the mosaic images to generate corresponding different feature maps for face detection.

16. A computer program product comprising at least one non-transitory computer-readable storage medium having computer-executable program code instructions stored therein, the computer-executable program code instruction comprising:

program code instructions for generating multiple mosaic images from an input image, each of the multiple mosaic images having a different scale;

program code instructions for employing a plurality of differently oriented edge detectors to perform edge detection on the multiple mosaic images including combining edges of the multiple mosaic images having the different scales; and

program code instructions for performing face detection in regions corresponding to the differently oriented edge detectors based on respective feature maps produced by the differently oriented edge detectors.

17. The computer program product of claim 16 , further comprising program code instructions for performing a verification of a candidate face detected by the face detection.

18. The computer program product of claim 16 , wherein program code instructions for employing the plurality of differently oriented edge detectors include instructions for utilizing edge detectors oriented with equal amounts of angular separation between each edge detector and at least one other edge detector and oriented to cover three hundred and sixty degrees of possible in-plane rotation angles for candidate faces in respective swaths of equal size.

19. The computer program product of claim 16 , wherein program code instructions for performing face detection include instructions for employing an elastic face template having a width corresponding to a distance between eyes of a detected eye pair and a height corresponding to a distance between a line connecting the detected eye pair and a potential mouth of a candidate face.

20. An apparatus comprising:

means for generating multiple mosaic images from an input image, each of the multiple mosaic images having a different scale;

means for employing a plurality of differently oriented edge detectors to perform edge detection on the multiple mosaic images including combining edges of the multiple mosaic images having the different scales; and

means for performing face detection in regions corresponding to the differently oriented edge detectors based on respective feature maps produced by the differently oriented edge detectors.