Image recognition apparatus, image recognition method, and recording medium

Patent metadata

Field	Value
Publication number	US-2016358338-A1
Application number	US-201615171551-A
Country	US
Kind code	A1
Filing date	Jun 2, 2016
Priority date	Jun 5, 2015
Publication date	Dec 8, 2016
Grant date	—

Abstract

On the basis of subsidiary information associated with image data, an image for the image data is segmented into multiple subregions, and feature values are extracted for each of the subregions obtained through the segmentation. The category for each of the subregions is determined on the basis of the extracted feature values.

First claim

What is claimed is:

1 . An image recognition apparatus comprising:
an acquiring unit configured to acquire image data and subsidiary information associated with the image data, wherein the subsidiary information includes at least information about an auto-exposure operation that is performed when an image for the image data is captured; a segmenting unit configured to segment the image for the acquired image data into a plurality of subregions on the basis of the acquired subsidiary information; an extracting unit configured to extract a feature value of each of the plurality of subregions obtained through the segmentation; and a determination unit configured to determine a category for each of the plurality of subregions on the basis of the extracted feature value.

2 . The image recognition apparatus according to claim 1 ,
wherein the information about the auto-exposure operation is information about a brightness value for each pixel in the image for the image data.

3 . The image recognition apparatus according to claim 1 ,
wherein the information about the auto-exposure operation is information about brightness values for each RGB channel of an image sensor that captures the image for the image data.

4 . The image recognition apparatus according to claim 1 ,
wherein the subsidiary information further includes information about automatic focusing and information about automatic white balance used when the image for the image data is captured.

5 . The image recognition apparatus according to claim 4 ,
wherein the information about automatic focusing is information about contrast values for each of a plurality of blocks which are obtained by segmenting the image for the image data.

6 . The image recognition apparatus according to claim 5 ,
wherein the plurality of blocks are set as regions identical to the plurality of subregions.

7 . The image recognition apparatus according to claim 4 ,
wherein the information about automatic white balance is information about a color temperature for each of a plurality of blocks which are obtained by segmenting the image for the image data.

8 . The image recognition apparatus according to claim 1 , further comprising:
a generating unit configured to generate the subsidiary information on the basis of the image data.

9 . The image recognition apparatus according to claim 1 ,
wherein the extracting unit extracts feature values for color, the subsidiary information, and texture as the feature value.

10 . The image recognition apparatus according to claim 1 , further comprising:
an intermediate-representation generating unit configured to convert the extracted feature value into a predetermined intermediate representation.

11 . The image recognition apparatus according to claim 1 , further comprising:
a scene recognizing unit configured to recognize a scene of the image for the image data on the basis of a global feature value of the image data, wherein the determination unit determines the category for each of the plurality of subregions on the basis of the recognized scene of the image.

12 . The image recognition apparatus according to claim 1 , further comprising:
a training unit configured to train a discriminator on the basis of training data and the subsidiary information of training image data, the discriminator determining the category for each of the plurality of subregions, the training data being data for which correct data of a category for each of a plurality of training subregions in the training image data is provided, wherein the determination unit determines the category for each of the plurality of subregions by using the trained discriminator.

13 . An image recognition method comprising:
acquiring image data and subsidiary information associated with the image data, wherein the subsidiary information includes at least information about an auto-exposure operation that is performed when an image for the image data is captured; segmenting the image for the acquired image data into a plurality of subregions on the basis of the acquired subsidiary information; extracting a feature value of each of the plurality of subregions obtained through the segmentation; and determining a category for each of the plurality of subregions on the basis of the extracted feature value.

14 . A non-transitory computer-readable recording medium that stores a program for causing a computer to function as the units of an image recognition apparatus comprising:
an acquiring unit configured to acquire image data and subsidiary information associated with the image data, wherein the subsidiary information includes at least information about an auto-exposure operation that is performed when an image for the image data is captured; a segmenting unit configured to segment the image for the acquired image data into a plurality of subregions on the basis of the acquired subsidiary information; an extracting unit configured to extract a feature value of each of the plurality of subregions obtained through the segmentation; and a determination unit configured to determine a category for each of the plurality of subregions on the basis of the extracted feature value.

Assignees

• Canon Kk

Inventors

• Tsunoda Takamasa

Classifications

• G06V10/7625

  Hierarchical techniques, i.e. dividing or merging patterns to obtain a tree-like representation; Dendograms · CPC title

• G06F18/24

  Classification techniques · CPC title

• G06V10/462 Primary

  Salient features, e.g. scale invariant feature transforms [SIFT] · CPC title

• G06F18/231

  Hierarchical techniques, i.e. dividing or merging pattern sets so as to obtain a dendrogram · CPC title

• G06K9/6267

  Physics · mapped topic