# US 10,121,085 B2 · Information processing apparatus and method of searching for similar data

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 10,121,085 B2
- Document ID: US10121085B2
- Status: Granted
- Kind: B2
- Application number: 15/425,129
- Filed: 2017-02-06
- Published: 2017-09-14
- Granted: 2018-11-06
- Earliest claimed: 2016-03-10
- Assignees: Fujitsu Limited
- Applicants: Fujitsu Limited
- Inventors: Yui Noma
- Web page: https://patentgrep.com/patent/US10121085B2

## CPC

- G06V 40/1365: {Matching; Classification}
- G06F 18/23: Clustering techniques
- G06F 18/28: Determining representative reference patterns, e.g. by averaging or distorting; Generating dictionaries
- G06V 10/462: {Salient features, e.g. scale invariant feature transforms [SIFT]}
- G06V 10/762: using clustering, e.g. of similar faces in social networks
- G06V 30/1914: {Determining representative reference patterns, e.g. averaging or distorting patterns; Generating dictionaries, e.g. user dictionaries}
- G06V 30/242: Division of the character sequences into groups prior to recognition; Selection of dictionaries
- G06V 40/10: Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V 40/14: Vascular patterns
- G06V: IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)

## Abstract

An information processing apparatus stores first and second registered feature data respectively expressing first and second features of registered data, generates first and second subject feature data respectively expressing the first and second features of subject data, calculates a first degree of dissimilarity between the first registered feature data and the first subject feature data using a first computational process that exhibits symmetry so that a computational result does not change when two input values are interchanged, calculates a second degree of dissimilarity between the second registered feature data and the second subject feature data using a second computational process that exhibits antisymmetry so that a computational result changes when the two input values are interchanged, and selects the registered data based on the first and second degrees of dissimilarity.

## Claims

### Claim 1

An information processing apparatus comprising:
a memory that stores therein first registered feature datasets respectively expressing a first feature of registered datasets and second registered feature datasets respectively expressing a second feature of the registered datasets; and
a processor configured to perform a procedure including:
receiving a subject dataset which is input to the information processing apparatus;
generating a first subject feature dataset expressing the first feature of the subject dataset and a second subject feature dataset expressing the second feature of the subject dataset;
calculating a first degree of dissimilarity that is a degree of dissimilarity between each of the first registered feature datasets and the first subject feature dataset using a first computational process that calculates a degree of dissimilarity between two input values and exhibits symmetry so that a computational result does not change when the two input values are interchanged;
calculating a second degree of dissimilarity that is a degree of dissimilarity between each of the second registered feature datasets and the second subject feature dataset using a second computational process that calculates a degree of dissimilarity between two input values and exhibits antisymmetry so that a computational result changes when the two input values are interchanged; and
selecting a first registered dataset from the registered datasets based on the first degree of dissimilarity and the second degree of dissimilarity, the first registered dataset being similar to the subject dataset.

### Claim 2 (depends on 1)

The information processing apparatus according to claim 1, wherein the processor determines that a degree of similarity between the subject dataset and each of the registered datasets is higher the smaller a sum of the first degree of dissimilarity and the second degree of dissimilarity.

### Claim 3 (depends on 1)

The information processing apparatus according to claim 1,
wherein the first registered feature datasets, the second registered feature datasets, the first subject feature dataset, and the second subject feature dataset are bit strings,
the first computational process is a computational process that counts a number of ones included in a result of a logical exclusive OR between one input value out of the two input values and another input value, and
the second computational process is a computational process that counts a number of ones included in a result of a logical AND between one input value out of the two input values and an inverted value produced by inverting another input value.

### Claim 4 (depends on 1)

The information processing apparatus according to claim 1,
wherein the first feature is a feature where, for a first dataset and a second dataset collected in a same environment, the higher a degree of similarity between the first feature of the first dataset and the first feature of the second dataset when the first dataset and the second dataset are compared, the lower the first degree of dissimilarity, and
the second feature is a feature where, for a third dataset and a fourth dataset collected in different environments, the higher a degree of similarity between the second feature of the third dataset and the second feature of the fourth dataset when the third dataset and the fourth dataset are compared, the lower the second degree of dissimilarity.

### Claim 5

A method of searching for similar data using a computer including a memory and a processor, comprising:
acquiring, by the processor, first registered feature datasets respectively expressing a first feature of registered datasets and second registered feature datasets respectively expressing a second feature of the registered datasets from the memory;
receiving, by the processor, a subject dataset which is input to the computer;
generating, by the processor, a first subject feature dataset expressing the first feature of the subject dataset and a second subject feature dataset expressing the second feature of the subject dataset;
calculating, by the processor, a first degree of dissimilarity between each of the first registered feature datasets and the first subject feature dataset using a first computational process that calculates a degree of dissimilarity between two input values and exhibits symmetry so that a computational result does not change when the two input values are interchanged;
calculating, by the processor, a second degree of dissimilarity between each of the second registered feature datasets and the second subject feature dataset using a second computational process that calculates a degree of dissimilarity between two input values and exhibits antisymmetry so that a computational result changes when the two input values are interchanged; and
selecting, by the processor, a first registered dataset from the registered datasets based on the first degree of dissimilarity and the second degree of dissimilarity, the first registered dataset being similar to the subject dataset.

### Claim 6

A non-transitory computer-readable storage medium storing a similar data search program that causes a computer including a memory and a processor to perform a procedure comprising:
acquiring, by the processor, first registered feature datasets respectively expressing a first feature of registered datasets and second registered feature datasets respectively expressing a second feature of the registered datasets from the memory;
receiving, by the processor, a subject dataset which is input to the computer;
generating, by the processor, a first subject feature dataset expressing the first feature of subject datasets and a second subject feature dataset expressing the second feature of the subject datasets;
calculating, by the processor, a first degree of dissimilarity between each of the first registered feature datasets and the first subject feature dataset using a first computational process that calculates a degree of dissimilarity between two input values and exhibits symmetry so that a computational result does not change when the two input values are interchanged;
calculating, by the processor, a second degree of dissimilarity between each of the second registered feature datasets and the second subject feature dataset using a second computational process that calculates a degree of dissimilarity between two input values and exhibits antisymmetry so that a computational result changes when the two input values are interchanged; and
selecting, by the processor, a first registered dataset from the registered datasets based on the first degree of dissimilarity and the second degree of dissimilarity, the first registered dataset being similar to the subject dataset.