APPARATUS, METHOD, AND PROGRAM FOR FACE FEATURE POINT DETECTION

• US 20080226175A1
• Filed: 03/12/2008
• Published: 09/18/2008
• Est. Priority Date: 03/13/2007
• Status: Active Grant

• Current Assignee: Aisin Seiki Co., Ltd. (Aisin Corp.), Toyota Jidosha Kabushiki Kaisha (Toyota Motor Corporation)
• Original Assignee: Aisin Seiki Co., Ltd. (Aisin Corp.), Toyota Jidosha Kabushiki Kaisha (Toyota Motor Corporation)
• Inventors: Suzuki, Tomoharu, Ninagawa, Yuji, Ohue, Kenichi, Adachi, Jun, Kojima, Shinichi, Uozumi, Shigeyasu, Takahashi, Kentaro, Yoshinaga, Yukihiko
• Granted Patent: US 7,936,926 B2
• Time in Patent Office: Days
• US Class Current: 382/190
• CPC Class Codes: G06V 40/168 Feature extraction; Face representation
• CPC Class Codes: G06V 40/18 Eye characteristics, e.g. of the iris

Abstract

A face feature point detection apparatus includes an image capturing device, an edge calculating unit calculating edge values indicating a luminance change in a direction, and a detection target determining unit scanning an edge image, which is created by arranging the edge values for corresponding pixels based on pixel arrangement of the face images, with an image window being an aggregation of selected pixels formed in a predetermined shape, the detection target determining unit determining a position of the image window having a largest weighted sum of weighted sums to be a detection position where a detection target is present, providing that the weighted sum is calculated by multiplying the edge value which corresponds to each pixel in the image window by a predetermined value defined on a per-pixel basis and adding up all products of the edge value and the predetermined value.

Claims

1. A face feature point detection apparatus, comprising:

an image capturing means capturing a face image;

an edge calculating means calculating edge values, each edge value indicating a luminance change in a direction, in the face image; and

a detection target determining means scanning an edge image, which is created by arranging the edge values for corresponding pixels calculated by the edge calculating means based on pixel arrangement of the face image, with an image window, the image window being an aggregation of selected pixels formed in a predetermined shape, the detection target determining means determining a position of the image window having a largest weighted sum of weighted sums to be a detection position where a detection target is present, providing that the weighted sum is calculated by multiplying the edge value which corresponds to each pixel in the image window by a predetermined value defined on a per-pixel basis and adding up all products of the edge value and the predetermined value.

2. A face feature point detection apparatus according to claim 1, wherein the image window is the aggregation of the selected pixels formed in the predetermined shape and comprising plural sub-windows, a constant positional relationship is maintained between the plural sub-windows while scanning the image, the detection target determining means determines the position of the image window having a largest total weighted so of total weighted sums to be the detection position that the detection target is present, the total weighted sum is a total of each weighted sum of the edge values correspond to each of the pixels in each sub-window.

3. A face feature point detection apparatus according to claim 2, wherein the edge calculating means creates a horizontal edge image by arranging horizontal edge values, each horizontal edge value indicating a luminance change in a horizontal direction, based on the pixel arrangement in the face image, the edge calculating means further creates a vertical edge image by arranging vertical edge values, each vertical edge value indicating a luminance change in a vertical direction, based on the pixel arrangement in the face image, the image window includes a horizontal edge window being an aggregation of selected pixels and a vertical edge window being an aggregation of selected pixels, the pixels of the horizontal and vertical edge windows are respectively selected to be formed in a shape of a portion to be scanned, a constant positional relationship is maintained between the horizontal edge window and the vertical edge window, the detection target determining means determines a position of the image window having a largest total weighted sum to be the detection position that the detection target is present, the total weighted sum is a total of a weighted sum of the horizontal edge values and a weighted sum of the vertical edge values, the horizontal edge value corresponds one-to-one with the pixels in the horizontal edge window, the vertical edge value corresponds one-to-one with the pixels in the vertical edge window.

4. A face feature point detection apparatus according to claim 1, further comprising:

a noise removing means removing an edge containing continuously clustered pixels whose number is less tan a predetermined value, providing that each pixel has the edge value whose absolute value is larger tan or equal to a predetermined threshold value.

5. A face feature point detection apparatus according to claim 3, wherein an eye is the detection target the vertical edge window corresponds to a vertical edge of an eyelid, and the horizontal edge window corresponds to at least one of a horizontal edge of an inner corner of the eye and a horizontal edge of an outer corner of the eye.

6. A face feature point detection apparatus according to claim 5, wherein the horizontal edge windows, corresponding to the inner and outer corners of the eye, are located at both sides of the vertical edge window at a level lower than the vertical edge window.

7. A face feature point detection apparatus according to claim 2, further comprising;

a noise removing means removing an edge containing continuously clustered pixels whose number is less than a predetermined value, providing that each pixel has the edge value whose absolute value is larger than a predetermined threshold value.

8. A face feature point detection apparatus according to claim 3, further comprising:

a noise removing means removing an edge containing continuously clustered pixels whose number is less than a predetermined value, providing that each pixel has the edge value whose absolute value is larger than a predetermined threshold value.

9. A face feature point detection apparatus according to claim 5, further comprising:

a noise removing means removing an edge containing continuously clustered pixels whose number is less than a predetermined value, providing that each pixel has the edge value whose absolute value is larger than a predetermined threshold value.

10. A face feature point detection apparatus according to claim 6, further comprising:

a noise removing means removing an edge containing continuously clustered pixels whose number is less than a predetermined value, providing that each pixel has the edge value whose absolute value is larger than a predetermined threshold value.

11. A face feature point detection method, comprising:

an edge calculating step calculating edge values, each edge value indicating a luminance change in a direction, in a face image; and

a detection target determining step scanning an edge image, which is created by arranging the edge values for corresponding pixels calculated by the edge calculating step based on pixel arrangement of the face image, with an image window, the image window being an aggregation of selected pixels formed in a predetermined shape, the detection target determining step determining a position of the image window having a largest weighted sum of weighted sums to be a detection position where a detection target is present, providing that the weighted sum is calculated by multiplying the edge value which corresponds to each pixel in the image window by a predetermined value defined on a per-pixel basis and adding up all products of the edge value and the predetermined value.

12. A program instructing a computer to function as;

an edge calculating means calculating edge values each edge value indicating a luminance change in a direction, in an face image; and

a detection target determining means scanning an edge image, which is created by arranging the edge values for corresponding pixels calculated by the edge calculating means based on pixel arrangement of the face image, with an image, window, the image window is an aggregation of selected pixels formed in a predetermined shape, the detection target determining means determining a position of the image window having a largest weighted sum of weighted sums to be a detection position where a detection target is present, providing that the weighted sum is calculated by multiplying be edge value which corresponds to each pixel in the image window by a predetermined value defined on a per-pixel basis and adding up all products of the edge value and the predetermined value.