Patents by Inventor Jerubbaal John Luke

Jerubbaal John Luke has filed for patents to protect the following inventions. This listing includes patent applications that are pending as well as patents that have already been granted by the United States Patent and Trademark Office (USPTO).

  • System and a method for semantic level image retrieval
    Patent number: 10949702
    Abstract: A system and method for retrieval of similar images related to query images is provided. The query images are pre-processed for noise removal by selecting filtering technique based on noise variance estimation in each query image with respect to pre-set noise variance threshold value. The pre-processed query images are pre-classified for determining class one image identifier. Image types are generated from pre-processed query images for determining class two image identifier. Features are extracted from pre-classified query images based on class one image identifier and from generated images based on class two image identifier. The images similar to query images are retrieved which have features similar to extracted features of pre-classified query images and generated images. The retrieved similar images are ranked for determining most similar images with respect to query images. Similarity between query images and retrieved similar images is analyzed for re-ranking retrieved similar images.
    Type: Grant
    Filed: June 13, 2019
    Date of Patent: March 16, 2021
    Assignee: COGNIZANT TECHNOLOGY SOLUTIONS INDIA PVT. LTD.
    Inventors: Rajkumar Joseph, Venugopal Gundimeda, Jerubbaal John Luke, Mahesh Balaji
  • SYSTEM AND A METHOD FOR SEMANTIC LEVEL IMAGE RETRIEVAL
    Publication number: 20200334486
    Abstract: A system and method for retrieval of similar images related to query images is provided. The query images are pre-processed for noise removal by selecting filtering technique based on noise variance estimation in each query image with respect to pre-set noise variance threshold value. The pre-processed query images are pre-classified for determining class one image identifier. Image types are generated from pre-processed query images for determining class two image identifier. Features are extracted from pre-classified query images based on class one image identifier and from generated images based on class two image identifier. The images similar to query images are retrieved which have features similar to extracted features of pre-classified query images and generated images. The retrieved similar images are ranked for determining most similar images with respect to query images. Similarity between query images and retrieved similar images is analyzed for re-ranking retrieved similar images.
    Type: Application
    Filed: June 13, 2019
    Publication date: October 22, 2020
    Inventors: Rajkumar Joseph, Venugopal Gundimeda, Jerubbaal John Luke, Mahesh Balaji