Method of detecting pedestrians using integral image of feature information, computer readable recording medium, and terminal device for the same
View Patent ↗Provided is a method of detecting a pedestrian through feature information prediction based on an integral image. The method includes extracting a feature image of an obtained original image; generating an integral image of the feature image; predicting an integral image of an enlarged/reduced feature image using the integral image of the feature image and an image enlargement/reduction scale value; and detecting a pedestrian using an integral image of the predicted enlarged/reduced feature image. Accordingly, it is possible to decrease a time taken for a feature extracting operation, and perform pedestrian detection efficiently.
1. A method of detecting pedestrians, the method comprising:
extracting a first feature image of a first image;
generating a first integral image of the first feature image;
calculating and predicting a second integral image of a second feature image using a scale value and the first integral image; and
detecting a pedestrian using the second integral image,
wherein the second feature image is either enlarged or reduced in size compared to the first feature image, and wherein the scale value is a size ratio between the first feature image and the second feature image.
2. The method according to claim 1 , wherein the scale value is k, and wherein the calculating and predicting are performed that a value of a point in the second integral image is calculated by multiplying k 2 to a corresponding point in the first integral image.
3. The method according to claim 1 , wherein the calculating and predicting calculate the second integral image by the following equation:
II R ( x, y )= k 2 II ( x/k, y/k )
wherein k is the scale value, wherein II is the first integral image, and wherein II R is the second integral image.
4. The method according to claim 1 , wherein the extracting is performed to extract each of channel features of the first image, and wherein the generating is performed to generate an integral image to each of the channel features of the first image.
5. The method according to claim 1 , wherein the method further comprises extracting a region of interest from the first image, and
wherein, when the region of interest is greater than a size of a pedestrian image in the first image, the pedestrian image is enlarged, and when the region of interest is smaller than the size of the pedestrian image in the first image, the first image is enlarged.
6. The method according to claim 1 , wherein the method further comprises recognizing and deciding the pedestrian using the second integral image.
7. A computer readable recording medium recording a computer program for performing a method of detecting pedestrians, the method comprising:
extracting a first feature image of a first image;
generating a first integral image of the first feature image;
calculating and predicting a second integral image of a second feature image using a scale value and the first integral image; and
detecting a pedestrian using the second integral image,
wherein the second feature image is either enlarged or reduced in size compared to the first feature image, and wherein the scale value is a size ratio between the first feature image and the second feature image.
8. The computer readable recording medium recording a computer program according to claim 7 , wherein the scale value is k, and wherein the calculating and predicting are performed that a value of a point in the second integral image is calculated by multiplying k 2 to a corresponding point in the first integral image.
9. The computer readable recording medium recording a computer program according to claim 7 , wherein the calculating and predicting calculate the second integral image by the following equation:
II R ( x, y )= k 2 II ( x/k, y/k )
wherein k is the scale value, wherein II is the first integral image, and wherein II R is the second integral image.
10. The computer readable recording medium recording a computer program according to claim 7 , wherein the extracting is performed to extract each of channel features of the first image, and wherein the generating is performed to generate an integral image to each of the channel features of the first image.
11. The computer readable recording medium recording a computer program according to claim 7 , wherein the method further comprises extracting a region of interest from the first image, and
wherein, when the region of interest is greater than a size of a pedestrian image in the first image, the pedestrian image is enlarged, and when the region of interest is smaller than the size of the pedestrian image in the first image, the first image is enlarged.
12. The computer readable recording medium recording a computer program according to claim 7 , wherein the method further comprises recognizing and deciding the pedestrian using the second integral image.