Method and system for detecting nationality of a financial document from layout of an input image of the financial document

Patent metadata

Field	Value
Publication number	US-10559156-B2
Application number	US-201715810405-A
Country	US
Kind code	B2
Filing date	Nov 13, 2017
Priority date	Nov 13, 2017
Publication date	Feb 11, 2020
Grant date	Feb 11, 2020

Abstract

The present subject matter is related in general to image processing that discloses a method for detecting nationality and layout of an input image. A nationality detection system retrieves predefined Financial Document (FD) images based on extracted features of an input image from a database and performs template matching of each predefined FD image with the input image to compute a first layout relevance score for each predefined FD image. Therefore, complexity of performing template matching with every predefined image in the database is eliminated, thereby increasing the processing speed. The nationality detection system detects the nationality and layout of the input image based on highest first layout relevance score if it is greater than or equal to predefined threshold value. Else, the nationality and layout of the input image are detected based on a nationality-based relevance score computed using the first layout relevance score of each predefined FD image.

First claim

The invention claimed is:

1. An image processing method comprising:
receiving, by a nationality detection system, an input image of a financial document (FD);
retrieving, by the nationality detection system, from a database associated with the nationality detection system, one or more predefined FD images based on one or more extracted features of the input image, wherein each of the one or more predefined FD images is associated with a predefined nationality;
performing, by the nationality detection system, one or more pre-processing steps on each of the one or more predefined FD images to obtain a first layout relevance score for each of the one or more predefined FD images;
performing, by the nationality detection system, when a highest first layout relevance score, among the first layout relevance score for each of the one or more predefined FD images, is less than a predefined threshold value, steps of:
for each predefined nationality associated with each of the one or more predefined FD images, computing, by the nationality detection system, a nationality-based relevance score based on the first layout relevance score of each of the one or more predefined FD images;
comparing, by the nationality detection system, a layout of the input image with a predefined layout of each of the one or more predefined FD images corresponding to a predefined nationality having a highest nationality-based relevance score;
for each of the one or more predefined FD images of the predefined nationality having the highest nationality-based relevance score, computing, by the nationality detection system, a second layout relevance score based on the comparison;
obtaining a highest second layout relevance score among the second layout relevance score of each of the one or more predefined FD images; and
detecting, by the nationality detection system, a nationality and the layout of the input image based on the predefined nationality and the predefined layout of a predefined FD image having the highest second layout relevance score, when the highest second layout relevance score is greater than or equal to the predefined threshold value,

repeating, by the nationality detection system, when the highest second layout relevance score is less than the predefined threshold value, steps of comparing, computing and detecting for a predefined nationalities with subsequent highest nationality-based relevance score consecutively until a subsequent highest second layout relevance score corresponding to one of the predefined nationalities with the subsequent highest nationality-based relevance score is greater than or equal to the predefined threshold value; and
performing, by the nationality detection system, when the subsequent highest second layout relevance score corresponding to the predefined nationalities is less than the predefined threshold value, the steps of:
assigning the predefined nationality having the highest nationality-based relevance score as a nationality of the input image; and
uploading the layout of the input image as a new layout associated with the corresponding nationality in the database.

2. The method as claimed in claim 1 further comprises:
detecting, by the nationality detection system, the nationality and the layout of the input image as the predefined nationality and the predefined layout of the predefined FD image corresponding to the highest first layout relevance score, when the highest first layout relevance score is greater than or equal to the predefined threshold value.

3. The method as claimed in claim 1 further comprises:
performing one or more operations of a financial institution by the nationality detection system upon detecting the nationality and the layout of the input image.

4. The method as claimed in claim 3 , wherein the one or more operations of the financial institution are at least one of segregating the FDs, counting the FDs, identifying denomination of the FDs and eliminating soiled FDs.

5. The method as claimed in claim 1 , wherein the one or more pre-processing steps comprises:
comparing, by the nationality detection system, the layout of the input image with the predefined layout corresponding to each of the one or more predefined FD images retrieved from the database;
for each of the one or more predefined FD images, computing, by the nationality detection system, the first layout relevance score based on the comparison; and
obtaining the highest first layout relevance score among the first layout relevance score of each of the one or more predefined FD images.

6. The method as claimed in claim 1 , wherein the one or more extracted features comprises at least one of color histogram, color values, and edge detection.

7. The method as claimed in claim 1 , wherein the input image is at least one of a complete image of the FD or a partial image of the FD.

8. The method as claimed in claim 1 , wherein the predefined layout of each of the one or more predefined FD images is associated with a predefined nationality.

9. A nationality detection system comprising:
a processor; and
a memory ( 113 ) communicatively coupled to the processor, wherein the memory stores processor-executable instructions, which, on execution, causes the processor to:
receive an input image of a financial document (FD);
retrieve, from a database associated with the nationality detection system, one or more predefined FD images based on one or more extracted features of the input image, wherein each of the one or more predefined FD images is associated with a predefined nationality;
perform one or more pre-processing steps on each of the one or more predefined FD images to obtain a first layout relevance score for each of the one or more predefined FD images;
perform when a highest first layout relevance score, among the first layout relevance score for each of the one or more predefined FD images, is less than a predefined threshold value, steps of:
for each predefined nationality associated with each of the one or more predefined FD images, computing a nationality-based relevance score based on the first layout relevance score of each of the one or more predefined FD images;
comparing a layout of the input image with a predefined layout of each of the one or more predefined FD images corresponding to a predefined nationality having a highest nationality-based relevance score;

for each of the one or more predefined FD images of the predefined nationality having the highest nationality-based relevance score,
computing a second layout relevance score based on the comparison;
obtaining a highest second layout relevance score among the second layout relevance score of each of the one or more predefined FD images; and
detecting a nationality and the layout of the input image based on the predefined nationality and the predefined layout of a predefined FD image having the highest second layout relevance score, when the highest second layout relevance score is greater than or equal to the predefined threshold value;

repeat, when the highest second layout relevance score is less than the predefined threshold value, steps of
comparing, computing and detecting for predefined nationalities with subsequent highest nationality-based relevance score consecutively until a subsequent highest second layout relevance score corresponding to one of the predefined nationalities with the subsequent highest nationality-based relevance score is greater than or equal to the predefined threshold value;

perform, when the subsequent highest second layout relevance score corresponding to the predefined nationalities is less than the predefined threshold value, the steps of:
assigning the predefined nationality having the highest na

Assignees

• Toshiba Kk
• Toshiba Infrastructure Systems & Solutions Corp

Inventors

• Sukumar Arya
• Joseph Rajkumar
• TIRUVEEDHULA Vijaya Kumar

Classifications

• G06V30/412

  Layout analysis of documents structured with printed lines or input boxes, e.g. business forms or tables · CPC title

• G06V10/40

  Extraction of image or video features · CPC title

• G06F16/90335

  Query processing · CPC title

• G07D7/206 Primary

  Matching template patterns · CPC title

• G06K9/00483

  Physics · mapped topic