Determining quality of an image or video using a distortion classifier
View Patent ↗Techniques and structures are disclosed in which one or more distortion categories are identified for an image or video, and a quality of the image or video is determined based on the one or more distortion categories. The image or video may be of a natural scene, and may be of unknown provenance. Identifying a distortion category and/or determining a quality may be performed without any corresponding reference (e.g., undistorted) image or video. Identifying a distortion category may be performed using a distortion classifier. Quality may be determined with respect to a plurality of human opinion scores that correspond to a particular distortion category to which an image or video of unknown provenance is identified as belonging. Various statistical methods may be used in performing said identifying and said determining, including use of generalized Gaussian distribution density models and natural scene statistics.
1. An apparatus, comprising:
a processor; and
a storage device having instructions stored thereon that are executable by the processor to cause the apparatus to perform operations including:
predicting one or more distortion categories from a plurality of distortion categories as having been previously applied to a first one or more image frames, wherein the predicting is based on a comparison of a plurality of image feature scores for the first one or more image frames with a plurality of image feature score ranges that correspond to one or more known distortion categories, wherein the plurality of image feature scores for the first one or more image frames are derived from a natural scene statistics model; and
determining a quality of the first one or more image frames based on the predicted one or more distortion categories, wherein the determining is based on a plurality of human-measured quality scores for a plurality of second one or more image frames, wherein each of the plurality of second one or more image frames are classified as being in at least one of the predicted one or more distortion categories, and wherein the determining includes a comparison of one or more the plurality of image feature scores for the first one or more image frames with other feature scores for the plurality of second one or more image frames.
2. The apparatus of claim 1 , wherein the predicting includes, for a particular distortion category of the plurality of distortion categories, assessing a probability of the particular distortion category being applicable to the first one or more image frames.
3. The apparatus of claim 1 , wherein reference image frames for the first one or more image frames are not available to the apparatus in performing the predicting identifying and the determining.
4. The apparatus of claim 1 , wherein the operations include receiving an upload of the one or more image frames from a user, and wherein the predicting and the determining are performed automatically in response to the receiving the upload.
5. The apparatus of claim 4 , wherein the apparatus includes one or more computer servers configured to couple to a network, and wherein the one or more computer servers are configured to perform one or more actions in response to the upload of the one or more image frames from the user.
6. A computer-readable storage medium having instructions stored thereon that are executable by a computer system to cause the computing device to perform operations comprising:
predicting one or more distortion categories from a plurality of distortion categories as having been previously applied to a first one or more image frames, wherein the predicting is based on a comparison of a plurality of image feature scores for the one or more image frames with a plurality of image feature score ranges that correspond to one or more known distortion categories; and
determining a quality of the first one or more image frames based on the predicted one or more distortion categories, wherein said determining is based on a plurality of human-measured quality scores for a plurality of second one or more image frames, wherein each of the plurality of second one or more image frames are classified as being in at least one of the identified one or more distortion categories, and wherein the determining includes a comparison of one or more of the plurality of image feature scores for the first one or more image frames with other feature scores for the plurality of second one or more image frames.
7. The computer-readable storage medium of claim 6 , wherein the predicting one or more distortion categories is performed using a distortion classifier for natural scene statistics.
8. The computer-readable storage medium of claim 6 , wherein at least one of the plurality of image feature scores is generated by applying a discrete cosine transform to the first one or more image frames.
9. The computer-readable storage medium of claim 6 , wherein reference image frames for the first one or more image frames are not available to the computer system in performing the predicting and the determining.
10. The computer-readable storage medium of claim 6 , wherein the instructions are executable by the computer system to cause the computing device to:
receive a plurality of uploaded one or more image frames from a plurality of users via a web interface; and
automatically determine quality for the plurality of uploaded one or more image frames.
11. The computer-readable storage medium of claim 10 , wherein the instructions are executable by the computer system to cause the computing device to, in response to automatically determining quality for each of the plurality of uploaded one or more image frames, automatically perform one or more actions on those one or more image frames.
12. A method, comprising:
a computer system predicting one or more distortion categories from a plurality of distortion categories as having been previously applied to a first one or more image frames, wherein the predicting is based on a comparison of a plurality of image feature scores for the first one or more image frames with a plurality of image feature score ranges that correspond to one or more known distortion categories; and
the computer system determining a quality of the first one or more image frames based on the predicted one or more distortion categories, wherein said determining is based on a plurality of human-measured quality scores for a plurality of second one or more image frames, wherein each of the plurality of second one or more image frames are classified as being in at least one of the identified one or more distortion categories, and wherein the determining includes a comparison of one or more of the plurality of image feature scores for the first one or more image frames with other feature scores for the plurality of second one or more image frames.
13. The method of claim 12 , wherein the predicting one or more distortion categories is performed using a distortion classifier for natural scene statistics.
14. The method of claim 12 , wherein at least one of the plurality of image feature scores is generated by applying a discrete cosine transform to the first one or more image frames.
15. The method of claim 12 , wherein reference image frames for the first one or more image frames are not available to the computer system in performing the identifying and the determining.
16. The method of claim 12 , wherein the predicting and the determining are performed in response to receiving an upload of the first one or more image frames from a user.
17. The method of claim 12 , wherein determining a quality of the first one or more image frames includes:
mapping from a multidimensional space to a quality score, wherein the multidimensional space is defined by the plurality of image feature score ranges.
18. The method of claim 12 , further comprising the computer system performing one or more corrections on the first one or more image frames, wherein performing the one or more corrections is in response to determining the quality of the one or more image frames and is based on the predicted one or more distortion categories.
19. The method of claim 12 , further comprising the computer system:
receiving a plurality of uploaded one or more image frames from a plurality of users via a web interface; and
automatically determining quality for the plurality of uploaded one or more image frames.
20. The method of claim 19 , further comprising the computer system, in response to automatically determining quality for each of the plurality of uploaded one or more image frames, automatically performing one or more actions on those one or more image frames.
21. The method of claim 20 , wherein the one or more actions include classifying individual ones of the plurality of uploaded one or more image frames as being of an acceptable level of quality.