Recognizing blank and nearly blank images
View Patent ↗A facility is provided for recognizing blank and nearly blank images. In various embodiments, the facility receives an indication of an image, creates an analysis image based on the received image by detecting edges, determines an edge ratio for the analysis image, determines whether the edge ratio exceeds a threshold and, when the edge ratio does not exceed the threshold, identifies the received image as a blank image. The facility can include an edge detector component that creates an analysis image based on a selected image by detecting edges in the selected image, a calculator component that calculates an edge ratio for the analysis image, and an analysis component that determines whether the calculated edge ratio exceeds a threshold.
1. A method performed by a computing device for recognizing a blank image, comprising:
receiving an indication of a video from which to select an image;
selecting an image from the indicated video;
creating an analysis image based on the selected image by detecting edges in the selected image;
determining an edge ratio for the analysis image;
determining whether the determining edge ratio exceeds a threshold; and
when the determined edge ratio does not exceed the threshold, identifying the selected image as a blank image.
2. The method of claim 1 wherein the selecting includes invoking a scene detection algorithm to select the image.
3. The method of claim 1 wherein a convolution algorithm detects edges.
4. The method of claim 1 further comprising selecting another image when the selected image is identified as a blank image.
5. The method of claim 1 wherein the blank image is nearly blank.
6. The method of claim 1 wherein the image is blank when it is mostly a solid color.
7. The method of claim 1 further comprising receiving an indication of the threshold.
8. A computer-readable medium having computer-executable instructions that, when executed, perform a method of recognizing a blank image, the method comprising:
receiving an indication of an image;
creating an analysis image based on the received image by detecting edges;
determining a ratio for the analysis image;
determining whether the determined ratio exceeds a threshold; and
when the determined ratio does not exceed the threshold, identifying the received image as a blank image.
9. The computer-readable medium of claim 8 wherein a convolution algorithm detects edges.
10. The computer-readable medium of claim 8 wherein the blank image is nearly blank.
11. The computer-readable medium of claim 8 wherein the image is blank when it is mostly a solid color.
12. The computer-readable medium of claim 8 further comprising receiving an indication of the threshold.
13. A system for recognizing a blank image, comprising:
an edge detector component that creates an analysis image based on a selected image by detecting edges in the selected image;
a calculator component that calculates an edge ratio for the analysis image; and
an analysis component that determines whether the calculated edge ratio exceeds a threshold, and when the calculated edge ratio does not exceed the threshold, identifies the selected image as a blank image.
14. The system of claim 13 wherein the edge detector component employs a convolution algorithm to detect edges.
15. The system of claim 13 wherein the calculator component calculates an edge ratio by calculating a quotient of a sum of non-black pixels in the analysis image divided by a sum of black pixels in the analysis image.
16. The system of claim 13 wherein the threshold is an edge ratio threshold.
17. The system of claim 13 wherein the image is blank when it is mostly a solid color.
18. The system of claim 13 wherein the analysis component receives an indication of the threshold.
19. The system of claim 13 wherein the threshold is 0.023.
20. The system of claim 13 wherein the edge detector component creates an analysis image by employing multiple edge detection algorithms.