Apparatus, monitoring system, method, and computer-readable medium
Provided is an apparatus comprising: an image acquisition unit configured to acquire a captured image; a compression unit configured to compress a captured image to generate compressed data; a reproduction unit configured to generate, from the compressed data, a reproduced image that reproduces the captured image; an evaluation acquisition unit configured to acquire an evaluation corresponding to a degree of approximation between the reproduced image and the captured image; and a learning processing unit configured to perform learning processing of a model configured to output, in response to an input of a new captured image, a compression parameter value to be applied in compression of the captured image, by using learning data including the evaluation, a captured image corresponding to the evaluation, and a compression parameter value applied in compression of the captured image.
1. An apparatus comprising:
an image acquisition unit configured to acquire a captured image;
a compression unit configured to perform compression of a captured image using at least one compression parameter having a compression parameter value to compress the captured image to generate compressed data;
a reproduction unit configured to generate, from the compressed data, a reproduced image that reproduces the captured image;
an evaluation acquisition unit configured to acquire an evaluation corresponding to a degree of approximation between the reproduced image and the captured image; and
a learning processing unit configured to perform learning processing of a model configured to output, in response to an input of a new captured image, a compression parameter value to be applied in compression of the new captured image, by using learning data including the evaluation, a captured image corresponding to the evaluation, and a compression parameter value applied in compression of the captured image corresponding to the evaluation, wherein
the image acquisition unit is configured to acquire a captured image captured under other image capturing condition different from a reference image capturing condition,
the compression unit is configured to generate, as the compressed data, an image for which compression has been performed by applying an image effect corresponding to the reference image capturing condition to the captured image, and
the reproduction unit is configured to generate, as the reproduced image, an image for which reproduction has been performed by applying an image effect corresponding to the other image capturing condition to an image of the compressed data.
2. The apparatus according to claim 1 , wherein
the captured image is a captured moving image,
the compression unit is configured to generate, as the compressed data, a moving image obtained by thinning out frames from the captured moving image, and
the reproduction unit is configured to generate, as the reproduced image, a moving image that reproduces the thinned-out frames.
3. The apparatus according to claim 1 , wherein
the compression unit is configured to irreversibly compress the captured image to generate the compressed data.
4. The apparatus according to claim 1 , wherein
the compression unit is configured to sequentially change the compression parameter value to sequentially generate compressed data corresponding to each compression parameter value.
5. The apparatus according to claim 1 , wherein
the learning processing unit is configured to perform learning processing of the model so that a compression parameter value to be applied in compression is between a compression parameter value of the compressed data corresponding to the evaluation, which is positive, and a compression parameter value of the compressed data corresponding to the evaluation, which is negative.
6. A method comprising:
acquiring a captured image;
performing compression of the captured image using at least one compression parameter having a compression parameter value to compress the captured image to generate compressed data;
generating, from the compressed data, a reproduced image that reproduces the captured image;
acquiring an evaluation corresponding to a degree of approximation between the reproduced image and the captured image;
performing learning processing of a model configured to output, in response to an input of a new captured image, a compression parameter value to be applied in compression of the new captured image, by using learning data including the evaluation, a captured image corresponding to the evaluation, and a compression parameter value applied in compression of the captured image corresponding to the evaluation;
acquiring a captured image captured under other image capturing condition different from a reference image capturing condition;
generating, as the compressed data, an image for which compression has been performed by applying an image effect corresponding to the reference image capturing condition to the captured image; and
generating, as the reproduced image, an image for which reproduction has been performed by applying an image effect corresponding to the other image capturing condition to an image of the compressed data.
7. A non-transitory, computer-readable medium having recorded thereon a program configured to cause a computer to function as:
an image acquisition unit configured to acquire a captured image;
a compression unit configured to perform compression of a captured image using at least one compression parameter having a compression parameter value to compress the captured image to generate compressed data;
a reproduction unit configured to generate, from the compressed data, a reproduced image that reproduces the captured image;
an evaluation acquisition unit configured to acquire an evaluation corresponding to a degree of approximation between the reproduced image and the captured image; and
a learning processing unit configured to perform learning processing of a model configured to output, in response to an input of a new captured image, a compression parameter value to be applied in compression of the new captured image, by using learning data including the evaluation, a captured image corresponding to the evaluation, and a compression parameter value applied in compression of the captured image corresponding to the evaluation, wherein
the image acquisition unit is configured to acquire a captured image captured under other image capturing condition different from a reference image capturing condition,
the compression unit is configured to generate, as the compressed data, an image for which compression has been performed by applying an image effect corresponding to the reference image capturing condition to the captured image, and
the reproduction unit is configured to generate, as the reproduced image, an image for which reproduction has been performed by applying an image effect corresponding to the other image capturing condition to an image of the compressed data.