Invention Application
- Patent Title: SELECTING TRAINING SYMBOLS FOR SYMBOL RECOGNITION
-
Application No.: US16965356Application Date: 2018-01-31
-
Publication No.: US20210124995A1Publication Date: 2021-04-29
- Inventor: Daniel Mas Montserrat , Qian Lin , Jan Allebach , Edward John Delp, III
- Applicant: Hewlett-Packard Development Company, L.P. , Purdue Research Foundation
- Applicant Address: US TX Spring; US IN West Lafayette
- Assignee: Hewlett-Packard Development Company, L.P.,Purdue Research Foundation
- Current Assignee: Hewlett-Packard Development Company, L.P.,Purdue Research Foundation
- Current Assignee Address: US TX Spring; US IN West Lafayette
- International Application: PCT/US2018/016211 WO 20180131
- Main IPC: G06K9/62
- IPC: G06K9/62 ; G06N3/08 ; G06F16/55 ; G06F16/58

Abstract:
A query is submitted to a search engine, where the query includes an identification of a symbol. A bounding box is generated in an unlabeled image returned by the search engine in response to the query. A confidence score is also generated that indicates a likelihood of the symbol being present in a portion of the unlabeled image enclosed by the bounding box. The unlabeled image is selected as a training image for training a system to recognize the symbol, when the confidence score is above a predefined threshold.
Information query