CONTENT SOFTENING OPTIMIZATION
A computer-implemented method comprising: receiving, as input, a plurality of images, each associated with a specified content category; generating, from each of the plurality of images, a set of transformed images by applying a series of non-photorealistic transformations having escalating transformation degrees, wherein each of the transformed images is labeled with a label indicating (i) the transformation degree applied thereto, and (ii) a content category associated therewith; obtaining, with respect to each of the set of transformed images, classification results assigned by a human annotator, wherein the classification results assign each of the transformed images in the set into one of a plurality of content categories; and calculating, for the human annotator, a classification score in each of the plurality of content categories, based, at least in part, on all of the classification results.
1 . A system comprising:
at least one hardware processor; and
a non-transitory computer-readable storage medium having stored thereon program instructions, the program instructions executable by the at least one hardware processor to:
receive, as input, a plurality of images, each associated with a specified content category,
generate, from each of said plurality of images, a set of transformed images by applying a series of non-photorealistic transformations having escalating transformation degrees, wherein each of said transformed images is labeled with a label indicating (i) said transformation degree applied thereto, and (ii) a content category associated therewith,
obtain, with respect to each of said set of transformed images, classification results assigned by a human annotator, wherein said classification results assign each of said transformed images in said set into one of a plurality of content categories, and
calculate, for said human annotator, a classification score in each of said plurality of content categories, based, at least in part, on all of said classification results.
2 . The system of claim 1 , wherein each of said images is one of: a single image, a series of images, a video segment, and video live streaming.
3 . The system of claim 1 , wherein, with respect to each of said images, each of said transformations represents at least one of: a softening of said image, a stylization of said image, an abstraction of said image, and a non-photorealistic rendering of said image.
4 . The system of claim 1 , wherein said plurality of transformation are selected from the group consisting of: color manipulation, line drawing, edge-preserving smoothing, contour transformations, edge detection and enhancement, tonal range modification, image-based artistic rendering.
5 . The system of claim 1 , wherein, with respect to each of said transformed images, said transformation degree corresponds to the level of recognizability of a content of said transformed image.
6 . The system of claim 5 , wherein said calculating of said classification score is based on the highest said transformation degree of a transformed image in said set that is correctly assigned to its associated specified content category.
7 . A computer-implemented method comprising:
receiving, as input, a plurality of images, each associated with a specified content category;
generating, from each of said plurality of images, a set of transformed images by applying a series of non-photorealistic transformations having escalating transformation degrees, wherein each of said transformed images is labeled with a label indicating (i) said transformation degree applied thereto, and (ii) a content category associated therewith;
obtaining, with respect to each of said set of transformed images, classification results assigned by a human annotator, wherein said classification results assign each of said transformed images in said set into one of a plurality of content categories; and
calculating, for said human annotator, a classification score in each of said plurality of content categories, based, at least in part, on all of said classification results.
8 . The computer-implemented method of claim 7 , wherein each of said images is one of: a single image, a series of images, a video segment, and video live streaming.
9 . The computer-implemented method of claim 7 , wherein, with respect to each of said images, each of said transformations represents at least one of: a softening of said image, a stylization of said image, an abstraction of said image, and a non-photorealistic rendering of said image.
10 . The computer-implemented method of claim 7 , wherein said plurality of transformation are selected from the group consisting of: color manipulation, line drawing, edge-preserving smoothing, contour transformations, edge detection and enhancement, tonal range modification, image-based artistic rendering.
11 . The computer-implemented method of claim 7 , wherein, with respect to each of said transformed images, said transformation degree corresponds to the level of recognizability of a content of said transformed image.
12 . The computer-implemented method of claim 11 , wherein said calculating of said classification score is based on the highest said transformation degree of a transformed image in said set that is correctly assigned to its associated specified content category.
13 . A computer program product comprising a non-transitory computer-readable storage medium having program instructions embodied therewith, the program instructions executable by at least one hardware processor to:
receive, as input, a plurality of images, each associated with a specified content category;
generate, from each of said plurality of images, a set of transformed images by applying a series of non-photorealistic transformations having escalating transformation degrees, wherein each of said transformed images is labeled with a label indicating (i) said transformation degree applied thereto, and (ii) a content category associated therewith;
obtain, with respect to each of said set of transformed images, classification results assigned by a human annotator, wherein said classification results assign each of said transformed images in said set into one of a plurality of content categories; and
calculate, for said human annotator, a classification score in each of said plurality of content categories, based, at least in part, on all of said classification results.
14 . The computer program product of claim 13 , wherein each of said images is one of: a single image, a series of images, a video segment, and video live streaming.
15 . The computer program product of claim 13 , wherein, with respect to each of said images, each of said transformations represents at least one of: a softening of said image, a stylization of said image, an abstraction of said image, and a non-photorealistic rendering of said image.
16 . The computer program product of claim 13 , wherein said plurality of transformation are selected from the group consisting of: color manipulation, line drawing, edge-preserving smoothing, contour transformations, edge detection and enhancement, tonal range modification, image-based artistic rendering.
17 . The computer program product of claim 13 , wherein, with respect to each of said transformed images, said transformation degree corresponds to the level of recognizability of a content of said transformed image.
18 . The computer program product of claim 17 , wherein said calculating of said classification score is based on the highest said transformation degree of a transformed image in said set that is correctly assigned to its associated specified content category.