Sorting system, recognition support apparatus, recognition support method, and recognition support program

Patent metadata

Field	Value
Publication number	US-2018200760-A1
Application number	US-201815861940-A
Country	US
Kind code	A1
Filing date	Jan 4, 2018
Priority date	Jan 13, 2017
Publication date	Jul 19, 2018
Grant date	—

Abstract

According to one embodiment, a sorting system has a delivery processing apparatus, a video coding terminal, and a recognition support apparatus. The recognition support apparatus has an information management portion, a correct solution derivation portion, and a machine learning portion. The information management portion acquires an image obtained by imaging a delivery object and keying information. The correct solution derivation portion derives respective correct solutions for a plurality of processings for the image obtained by imaging the delivery object, based on the keying information acquired by the information management portion. The machine learning portion performs machine learning using the respective correct solutions for the plurality of processings derived by the correct solution derivation portion, to adjust the plurality of processings.

First claim

1 . A sorting system comprising:
a delivery processing apparatus which performs an OCR processing including a plurality of processings for an image obtained by imaging a delivery object, and conveys the delivery object to a stacking portion corresponding to information read by the OCR processing; a video coding terminal which receives the image obtained by imaging the delivery object from the delivery processing apparatus, displays the received image, and transmits keying information inputted to the image obtained by imaging the delivery object to the delivery processing apparatus; and a recognition support apparatus wherein the recognition support apparatus further comprising, an information management portion which acquires the image obtained by imaging the delivery object and the keying information, a correct solution derivation portion which derives respective correct solutions for the plurality of processings for the image obtained by imaging the delivery object, based on the keying information acquired by the information management portion, and a machine learning portion which performs machine learning using the respective correct solutions for the plurality of processing derived by the correct solution derivation portion, to adjust the plurality of processings.

2 . The sorting system according to claim 1 , wherein:
the machine learning portion performs the machine learning by deep learning.

3 . The sorting system according to claim 1 , wherein:
the correct solution derivation portion derives the respective correct solutions for the plurality of processings for the image obtained by imaging the delivery object in order from a downstream side, with respect to a flow of the processing.

4 . A recognition support apparatus comprising:
an information management portion which acquires an image that is obtained by imaging a delivery object in a delivery processing apparatus and becomes an object of an OCR processing including a plurality of processings, and keying information inputted to the image; a correct solution derivation portion which derives respective correct solutions for the plurality of processings for the image obtained by imaging the delivery object, based on the keying information acquired by the information management portion; and a machine learning portion which performs machine learning using the respective correct solutions for the plurality of processing derived by the correct solution derivation portion, to adjust the plurality of processings.

5 . The recognition support apparatus according to claim 4 , wherein:
the machine learning portion performs the machine learning by deep learning.

6 . The recognition support apparatus according to claim 4 , wherein:
the correct solution derivation portion derives the respective correct solutions for the plurality of processings for the image obtained by imaging the delivery object in order from a downstream side, with respect to a flow of the processing.

7 . A recognition support method which a computer performs comprising:
acquiring an image that is obtained by imaging a delivery object in a delivery processing apparatus and becomes an object of an OCR processing including a plurality of processings, and keying information inputted to the image; deriving respective correct solutions for the plurality of processings for the image obtained by imaging the delivery object, based on the acquired keying information; and performing machine learning using the respective derived correct solutions for the plurality of processings, to adjust the plurality of processings.

8 . A recognition support program comprising:
making a computer acquire an image that is obtained by imaging a delivery object in a delivery processing apparatus and becomes an object of an OCR processing including a plurality of processings, and keying information inputted to the image; making the computer derive respective correct solutions for the plurality of processing for the image obtained by imaging the delivery object, based on the acquired keying information; and making the computer perform machine learning using the respective derived correct solutions for the plurality of processings, to adjust the plurality of processings.

9 . The sorting system according to claim 2 , wherein:
the correct solution derivation portion derives the respective correct solutions for the plurality of processings for the image obtained by imaging the delivery object in order from a downstream side, with respect to a flow of the processing.

10 . The recognition support apparatus according to claim 5 , wherein:
the correct solution derivation portion derives the respective correct solutions for the plurality of processings for the image obtained by imaging the delivery object in order from a downstream side, with respect to a flow of the processing.

Assignees

• Toshiba Kk
• Toshiba Infrastructure Systems & Solutions Corp

Inventors

• Ohkawa Yasuhiro
• Hamamura Tomoyuki
• Akagi Takuma

Classifications

• G06V30/18057

  Integrating the filters into a hierarchical structure, e.g. convolutional neural networks [CNN] · CPC title

• G06V30/127

  with the intervention of an operator · CPC title

• G06V10/82

  using neural networks · CPC title

• G06V30/10

  Character recognition · CPC title

• B07C3/14 Primary

  using light-responsive detecting means {(for information carriers G06K7/14)} · CPC title