Invention Grant
- Patent Title: Generating artificial images for use in neural networks
-
Application No.: US15900921Application Date: 2018-02-21
-
Publication No.: US10586129B2Publication Date: 2020-03-10
- Inventor: Marc Fiammante
- Applicant: INTERNATIONAL BUSINESS MACHINES CORPORATION
- Applicant Address: US NY Armonk
- Assignee: International Business Machines Corporation
- Current Assignee: International Business Machines Corporation
- Current Assignee Address: US NY Armonk
- Agent Tihon Poltavets
- Main IPC: G06K9/42
- IPC: G06K9/42 ; G06K9/62 ; G06K9/46

Abstract:
A method includes creating an input matrix of image data for a received input image and responsive to selecting a first image filter for generating an artificial image, the method applies a first filter computation for the first image filter to the input matrix of image data, where applying the first filter computation generates a first resulting matrix of image data. The method concatenates the first resulting matrix of image data to the input matrix of image data to generate a first concatenated matrix of image data. The method generates an artificial image suitable for use in neural networks based on the concatenated matrix of image data.
Public/Granted literature
- US20190258898A1 GENERATING ARTIFICIAL IMAGES FOR USE IN NEURAL NETWORKS Public/Granted day:2019-08-22
Information query