Apparatus and method that determines whether a pattern within the detection window is a subject based on characteristic amounts obtained from within a first region

• US 8,873,859 B2
• Filed: 09/01/2010
• Issued: 10/28/2014
• Est. Priority Date: 09/02/2009
• Status: Expired due to Fees

• Current Assignee: Canon Kabushiki Kaisha (Canon Inc.)
• Original Assignee: Canon Kabushiki Kaisha (Canon Inc.)
• Inventors: Yashiro, Satoshi
• Primary Examiner(s): Repko, Jason M
• Assistant Examiner(s): ROWLAND, JHERE M
• Application Number:
  US12/874,046
• Publication Number:
  US 20110052076A1
• Time in Patent Office: 1,518 Days
• Field of Search: None
• US Class Current: 382/190
• CPC Class Codes: G06F 18/25 Fusion techniques
• CPC Class Codes: G06V 10/751 Comparing pixel values or logical combinations thereof, or feature values having positional relevance, e.g. template matching
• CPC Class Codes: G06V 10/80 Fusion, i.e. combining data from various sources at the sensor level, preprocessing level, feature extraction level or classification level multimodal speaker identification or verification G10L17/10
• CPC Class Codes: G06V 40/10 Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands

Abstract

An image processing apparatus includes an image input unit configured to input an image, a scanning unit configured to scan a detection window on the input image, a first discrimination unit configured to determine whether a pattern within the detection window is a subject based on a plurality of characteristic amounts obtained from within a first region among a plurality of regions within the detection window, and a second discrimination unit configured to determine, if it is determined that the pattern is not the subject by the first discrimination unit, whether the pattern is the subject based on a plurality of characteristic amounts obtained from a second region in which a probability that occlusion of the subject occurs is higher than that in the first region among the plurality of regions. Accordingly, a subject can be detected efficiently, and omissions of detection can be reduced.

Claims

1. An image processing apparatus comprising:

a processor comprising;

an image input unit configured to input an image;

a scanning unit configured to scan the input image with a moving detection window on the input image;

a first discrimination unit configured to determine whether a pattern within the detection window contains a predetermined subject based on a plurality of characteristic amounts obtained from a first region corresponding to a specific part of the predetermined subject within the detection window; and

a second discrimination unit configured to determine, if it is not determined that the pattern contains the predetermined subject by the first discrimination unit, whether the pattern within the detection window contains the predetermined subject based on a plurality of characteristic amounts obtained from a second region corresponding to the whole of the predetermined subject within the detection window,

wherein if the pattern is determined to be the predetermined subject by the first discrimination unit a determination by the second discrimination unit is omitted for the pattern within the detection window,

the first region is smaller than the second region and is placed at a predetermined position within the second region, and

a probability of occurring occlusion in the first region is lower than that in the second region other than first region.

2. The image processing apparatus according to claim 1, wherein the first discrimination unit calculates subject reliability concerning whether the pattern within the detection window contains the predetermined subject based on the plurality of characteristic amounts obtained from within the first region, and determines that the pattern contains the predetermined subject if the subject reliability is equal to or greater than a threshold value.

3. The image processing apparatus according to claim 1, wherein the first region is a region of an upper body of a person, and the second region is a region of an entire body of the person.

4. The image processing apparatus according to claim 1, further comprising an integration unit configured to integrate the detection windows of a plurality of positions with respect to the identical pattern which is determined to contain the predetermined subject by the second discrimination unit based on positions of the respective detection windows and the subject reliability calculated based on a plurality of characteristic amounts obtained from within regions within the respective detection windows.

5. The image processing apparatus according to claim 1, further comprising a correction unit configured to correct the subject reliability, if the subject reliability calculated by the first discrimination unit based on the plurality of characteristic amounts obtained from within the first region is equal to or greater than a threshold value.

8. The image processing apparatus according to claim 1, wherein

the first discrimination unit calculates a first subject reliability concerning whether the pattern within the detection window contains the predetermined subject based on the plurality of characteristic amounts obtained from the first region corresponding to the specific part of the predetermined subject within the detection window;

the second discrimination unit calculates a second subject reliability concerning whether the pattern within the detection window contains the predetermined subject based on the plurality of characteristic amounts obtained from the second region corresponding to the whole of the predetermined subject within the detection window; and

the first subject reliability always exceeds a maximum value of the second subject reliability.

6. A method for subject discrimination performed by an image processing apparatus, the method comprising:

inputting an input image;

scanning the input image with a moving detection window on the input image;

determining whether a pattern within the detection window contains a predetermined subject based on a plurality of characteristic amounts obtained from a first region corresponding to a specific part of the predetermined subject within the detection window; and

if it is not determined that the pattern contains the predetermined subject, determining whether the pattern within the detection window contains the predetermined subject based on a plurality of characteristic amounts obtained from within a second region corresponding to the whole of the predetermined subject within the detection window,

if it is determined that the pattern is the subject then do not determine whether the pattern within the detection window is the subject from within the second region

the first region is smaller than the second region and is placed at a predetermined position within the second region, and

a probability of occurring occlusion in the first region is lower than that in the second region other than first region.

7. A non-transitory computer-readable storage medium that stores a program for causing a computer to function as,

an image input unit configured to input an image;

a scanning unit configured to scan the input image with a moving detection window on the input image;

a first discrimination unit configured to determine whether a pattern within the detection window contains a predetermined subject based on a plurality of characteristic amounts obtained from a first region corresponding to a specific part of the predetermined subject within the detection window; and

a second discrimination unit configured to determine, if it is not determined that the pattern contains the predetermined subject by the first discrimination unit, whether the pattern within the detection window contains the predetermined subject based on a plurality of characteristic amounts obtained from a second region corresponding to the whole of the predetermined subject within the detection window,

if it is determined that the pattern is the predetermined subject then do not determine whether the pattern within the detection window is the subject from within the second region,

the first region is smaller than the second region and is placed at a predetermined position within the second region, and

a probability of occurring occlusion in the first region is lower than that in the second region other than first region.