Artificially intelligent, machine learning-based, image enhancement, processing, improvement and feedback algorithms

Patent metadata

Field	Value
Publication number	US-11055822-B2
Application number	US-201916402409-A
Country	US
Kind code	B2
Filing date	May 3, 2019
Priority date	May 3, 2019
Publication date	Jul 6, 2021
Grant date	Jul 6, 2021

Abstract

Systems, methods, and computer program products leverage artificial intelligence, and machine learning to process image enhancements using image enhancement techniques and algorithms. Image enhancements are determined to be best suited for enhancing each image as a function of each images' calculated validation parameters by an analytics engine. The images are each categorized by the image quality as a function of the validation parameters. Images identified as having an improvement space are further processed by querying the images' validation parameters using a knowledge base comprising historical data describing past image enhancements and historical validation parameters to the current image. A matrix of recommended enhancements, along with a predicted success rate for improving the image quality is provided to a user interface. A user can select one or more enhancements to apply to the image(s) and further provide feedback to the knowledge base, further improving enhancement recommendations and success rates.

First claim

What is claimed is:

1. A computer-implemented method comprising the steps of:
receiving an image set comprising at least one image;
calculating validation parameters for each image of the image set;
categorizing each image of the image set based on the validation parameters, wherein the step of categorizing identifies a category of images in the image set that have an available improvement space for enhancement;
querying a knowledge base for recommended image enhancements based on the validation parameters for each image having the available improvement space;
generating a matrix comprising a set of one or more recommended image enhancements and a set of success rates for each enhancement recommended;
displaying on a user interface, the one or more recommended image enhancements, a predicted rate of success for applying the one or more recommended image enhancements, a predicted rate of improvement to the improvement space of each image, and a preview of each image of the image set after applying the one or more recommended image enhancements;
customizing the one or more recommended image enhancements displayed by the user interface by editing one or more values of the matrix to a desired value selected by the user;
applying the one or more recommended image enhancements selected by a user from the user interface, to at least one image of the image set having the available improvement space, including application of at least one customized image enhancement comprising the one or more values of the matrix edited by the user, creating at least one enhanced image;
outputting an enhanced image set, wherein the enhanced image set replaces at least one image of the image set with the at least one enhanced image;
receiving user feedback critiquing the enhanced image set; and
storing user feedback to the knowledge base, improving future recommended image enhancements for images comprising similar validation parameters.

2. The method of claim 1 , wherein the step of categorizing each image of the image set includes sorting each image of the image set into a category selected from the group consisting of an inferior quality image, a superior quality image and an average image with the available improvement space.

3. The method of claim 2 , further comprising the steps of:
marking to ignore each image of the image set categorized as the inferior quality image during the outputting of the enhanced image set; and
including each superior quality image, un-enhanced, within the enhanced image set.

4. The method of claim 1 , wherein the validation parameters for each image of the image set are selected from the group consisting of dots per inch (DPI), length and width of the image, a histogram of the image, pixel distribution, image rotation, image layers, content distribution, and a combination thereof.

5. The method of claim 1 , further comprising the steps of:
selecting, by the user, a first image enhancement from the matrix to apply to a first image of the image set; and
selecting, by the user, a second image enhancement from a second matrix comprising a second set of one or more recommended image enhancements to apply to a second image of the image set, and a second set of success rates for each enhancement proposed for the second image, wherein the first image enhancement is different from the second image enhancement.

6. The method of claim 1 , wherein the one or more image enhancements selected by the user are selected from the matrix comprising the set of one or more recommended image enhancements.

7. The method of claim 1 , wherein the one or more image enhancements selected by the user are manually selected by the user, and the image enhancements selected by the user are not part of the set of one or more recommended image enhancements described by the matrix.

8. A computer system comprising:
a processor; and
a computer-readable storage media coupled to the processor, wherein the computer-readable storage media contains program instructions executing a computer-implemented method comprising the steps of:
receiving an image set comprising at least one image,
calculating validation parameters for each image of the image set,
categorizing each image of the image set based on the validation parameters, wherein the step of categorizing identifies a category of images in the image set that have an available improvement space,
querying a knowledge base for recommended image enhancements based on the validation parameters for each image having the available improvement space,
generating a matrix comprising a set of one or more recommended image enhancements and a set of success rates for each enhancement recommended,
displaying on a user interface, the one or more recommended image enhancements, a predicted rate of success for applying the one or more recommended image enhancements, a predicted rate of improvement to the improvement space of each image, and a preview of each image of the image set after applying the one or more recommended image enhancements,
customizing the one or more recommended image enhancements displayed by the user interface by editing one or more values of the matrix to a desired value selected by the user,
applying the one or more recommended image enhancements selected by a user from the user interface, to at least one image of the image set having the available improvement space, including application of at least one customized image enhancement comprising the one or more values of the matrix edited by the user, creating at least one enhanced image,
outputting an enhanced image set, wherein the enhanced image set replaces at least one image of the image set with the at least one enhanced image,
receiving user feedback critiquing the enhanced image set, and
storing user feedback to the knowledge base, improving future recommended image enhancements for images comprising similar validation parameters.

9. The computer system of claim 8 , wherein the step of categorizing each image of the image set includes sorting each image of the image set into a category selected from the group consisting of an inferior quality image, a superior quality image and an average image with the available improvement space.

10. The computer system of claim 9 , further comprising the steps of:
marking to ignore each image of the image set categorized as the inferior quality image during the outputting of the enhanced image set; and
including each superior quality image, un-enhanced, within the enhanced image set.

11. The computer system of claim 8 , wherein the validation parameters for each image of the image set are selected from the group consisting of dots per inch (DPI), length and width of the image, a histogram of the image, pixel distribution, image rotation, image layers, content distribution, and a combination thereof.

12. The computer system of claim 8 , further comprising the steps of:
selecting, by the user, a first image enhancement from the matrix to apply to a first image of the image set; and
selecting, by the user, a second image enhancement from a second matrix comprising a second set of one or more recommended image enhancements to apply to a second image of the image set, and a second set of success rates for each enhancement proposed for the second image, wherein the first image enhancement is different from the second image enhancement.

13. The computer system of claim 8 , wherein the one or more image enhancements selected by the user are manually selected by the user, and the image enhancements selected by the user are not part of the set of one or more recommended image enhancements described by the matr

Assignees

• IBM

Inventors

• Ramanujam Karthick
• Gangadhar Jayanth
• Basavaraj Ullas M
• Shah Ankur Bharatkumar

Classifications

• G06V10/7784

  based on feedback from supervisors · CPC title

• G06F18/2178

  based on feedback of a supervisor · CPC title

• G06N3/09

  Supervised learning · CPC title

• G06N3/0464

  Convolutional networks [CNN, ConvNet] · CPC title

• G06N3/0895

  Weakly supervised learning, e.g. semi-supervised or self-supervised learning · CPC title