Image recognition apparatus, image recognition method, and recording medium
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.
1. An image recognition apparatus comprising:
a memory; and
a processor in communication with the memory, the processor configured to provide:
an acquiring unit configured to acquire image data of a plurality of color components for a captured image and an image capturing parameter used for capturing the image;
a generation unit configured to generate, based on the acquired image capturing parameter and the acquired plurality of color components, a plurality of evaluation maps, wherein each map represents evaluation values for a plurality of regions in an image represented by the acquired image data and a respective one of the plurality of color components, the evaluation values including at least an evaluation value about exposure that has been performed when the image was captured;
an extracting unit configured to extract a feature value from each of the plurality of regions in the image represented by the acquired image data; and
a determination unit configured to determine a category for each of the plurality of regions on a basis of the evaluation values for the color components and the extracted feature value.
2. The image recognition apparatus according to claim 1 ,
wherein the evaluation value about the exposure is an evaluation value 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 evaluation value about the exposure is an evaluation value about a 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 evaluation values further include an evaluation value about automatic focusing and an evaluation value about automatic white balance.
5. The image recognition apparatus according to claim 4 ,
wherein the evaluation value about automatic focusing is an evaluation value about a 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 subregions identical to the plurality of regions.
7. The image recognition apparatus according to claim 4 ,
wherein the evaluation value about automatic white balance is an evaluation value 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 ,
wherein the extracting unit extracts feature values for color and texture of the image as the feature value.
9. The image recognition apparatus according to claim 1 , wherein the processor is further configured to provide:
an intermediate-representation generating unit configured to convert the extracted feature value into a predetermined intermediate representation.
10. The image recognition apparatus according to claim 1 , wherein the processor is further configured to provide:
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 regions on the basis of the recognized scene of the image.
11. The image recognition apparatus according to claim 1 , wherein the processor is further configured to provide:
a training unit configured to train a discriminator on a basis of training data and the evaluation value of training image data, the discriminator determining the category for each of the plurality of regions, the training data being data for which correct data of a category for each of a plurality of training regions in the training image data is provided,
wherein the determination unit determines the category for each of the plurality of regions by using the trained discriminator.
12. The image recognition apparatus according to claim 1 ,
wherein each of the plurality of regions includes multiple pixels of the image.
13. The image recognition apparatus according to claim 1 ,
wherein, as the image capturing parameter, the acquiring unit acquires an aperture value at a timing of capturing the image, exposure time at a timing of capturing the image, and sensitivity of an image sensor at a timing of capturing the image.
14. The image recognition apparatus according to claim 1 ,
wherein the image data is RAW data.
15. The image recognition apparatus according to claim 1 , wherein the processor is further configured to provide:
a segmenting unit configured to segment the captured image into the plurality of regions on a basis of the evaluation value;
wherein, on the basis of the evaluation value and the feature value for each of the plurality of regions obtained through the segmentation, the determination unit determines the category for each of the plurality of regions.
16. An image recognition method comprising:
acquiring image data of a plurality of color components for a captured image and an image capturing parameter used for capturing the image;
generating, based on the acquired image capturing parameter and the acquired plurality of color components of the captured image, a plurality of evaluation maps, wherein each map represents evaluation values for a plurality of regions in an image represented by the acquired image data and a respective one of the plurality of color components, the evaluation values including at least an evaluation value about exposure that has been performed when the image was captured;
extracting a feature value from each of the plurality of regions in the image represented by the acquired image data; and
determining a category for each of the plurality of regions on a basis of the evaluation values for the color components and the extracted feature value.
17. A non-transitory computer-readable recording medium that stores a program for causing a computer to execute an image recognition method comprising:
acquiring image data of a plurality of color components for a captured image and an image capturing parameter used for capturing the image;
generating, based on the acquired image capturing parameter and the acquired plurality of color components of the captured image, a plurality of evaluation maps, wherein each map represents evaluation values for a plurality of regions in an image represented by the acquired image data and a respective one of the plurality of color components, the evaluation values including at least an evaluation value about exposure that has been performed when the image was captured;
extracting a feature value from each of the plurality of regions in the image represented by the acquired image data; and
determining a category for each of the plurality of regions on a basis of the evaluation values for the color components and the extracted feature value.