Method and a system for context based clustering of object

Patent metadata

Field	Value
Publication number	US-11537938-B2
Application number	US-201916370900-A
Country	US
Kind code	B2
Filing date	Mar 30, 2019
Priority date	Feb 15, 2019
Publication date	Dec 27, 2022
Grant date	Dec 27, 2022

Abstract

A method and a system are described for context based clustering of one or more objects. The method comprises receiving, by the object clustering system, receiving, by an object clustering system, an object clustering request for one or more objects associated with a plurality of contextual parameters, where the plurality of contextual parameters comprises one or more physical attributes and one or more non-physical attributes. It further includes tagging the one or more non-physical attributes respectively to the one or more physical attributes. It further includes identifying a common context from the one or more physical attributes associated with the one or more objects based on the tagging. It further includes mapping the one or more physical attributes to the one or more objects based on the common context. It then includes clustering the one or more objects based on the mapping.

First claim

We claim:

1. A method of context based clustering of objects, the method comprising:
receiving, by an object clustering system, an object clustering request for one or more objects associated with a plurality of contextual parameters, wherein the plurality of contextual parameters comprises one or more physical attributes and one or more non-physical attributes;
tagging, by the object clustering system, the one or more non-physical attributes respectively to the one or more physical attributes;
identifying, by the object clustering system, a common context from the one or more physical attributes associated with the one or more objects based on the tagging;
mapping, by the object clustering system, the one or more physical attributes to the one or more objects based on the common context;
determining, by the object clustering system, a mean value of the one or more physical attributes by generating a histogram, and a weightage is assigned to the one or more physical attributes;
computing, by the object clustering system, a variance value of the one or more physical attributes from the mean value of the histogram; and
clustering, by the object clustering system, the one or more objects based on the mapping when a distance of the variance value is lesser than a pre-defined distance with respect to the mean value.

2. The method as claimed in claim 1 , wherein the one or more physical attributes comprises at least one of colors, shapes, names, and places, and the one or more non-physical attributes comprises at least one of user emotions, user gestures, pronouns, tastes, and smells.

3. The method as claimed in claim 1 , wherein tagging the one or more non-physical attributes further comprises: identifying a correlation between the one or more non-physical attributes and the one or more physical attributes.

4. The method as claimed in claim 1 further comprising:
creating another cluster for the one or more objects if the distance between the variance value with respect to the mean value is more than the pre-defined distance.

5. The method as claimed in claim 1 further comprising:
learning to identify the or more objects adaptively, wherein the learning comprises:
training the object clustering system with multimodalities of inputs comprising at least one of documents, images, videos and audios, wherein the training further comprises identifying texts from the documents, one or more figures from the images, one or more figures from image frames of the videos and one or more audio signatures from the audios;
identifying the plurality of contextual parameters associated with the one or more objects;
learning a pattern over a historical usage of the clustering of the one or more objects based on the common context; and
incorporating user feedback on the clustering of the one or more objects associated with the common context.

6. An object clustering system for context based clustering of objects comprising:
a processor; and
a memory communicatively coupled to the processor, wherein the memory stores processor executable instructions, which on execution causes the processor to:
receive an object clustering request for one or more objects associated with a plurality of contextual parameters, wherein the plurality of contextual parameters comprises one or more physical attributes and one or more non-physical attributes;
tag the one or more non-physical attributes respectively to the one or more physical attributes;
identify a common context from the one or more physical attributes associated with the one or more objects based on the tagging;
map the one or more physical attributes to the one or more objects based on the common context;
determine a mean value of the one or more physical attributes by generating a histogram, and a weightage is assigned to the one or more physical attributes;
compute a variance value of the one or more physical attributes from the mean value of the histogram; and
cluster the one or more objects based on the mapping when a distance of the variance value is lesser than a pre-defined distance with respect to the mean value.

7. The object clustering of claim 6 , wherein the one or more physical attributes comprises at least one of colors, shapes, names, and places, and the one or more non-physical attributes comprises at least one of user emotions, user gestures, pronouns, tastes, smells and locations.

8. The object clustering of claim 6 , wherein tagging the one or more non-physical attributes further comprises: identifying a correlation between the one or more non-physical attributes and the one or more physical attributes.

9. The object clustering of claim 6 further comprising:
creating a new cluster for the one or more objects if the distance between the variance value with respect to the mean value is more than the pre-defined distance.

10. The object clustering of claim 6 further comprising:
learning to identify the or more objects adaptively, wherein the learning comprises:
training the object clustering system with multimodalities of inputs comprising at least one of documents, images, videos and audios, wherein the training further comprises identifying texts from the documents, one or more figures from the images, one or more figures from image frames of the videos and one or more audio signatures from the audios;
identifying the plurality of contextual parameters associated with the one or more objects;
learning a pattern over a historical usage of the clustering of the one or more objects based on the common context; and
incorporating user feedback on the clustering of the one or more objects associated with the common context.

11. A non-transitory computer-readable storage medium having stored thereon, a set of computer-executable instructions causing a computer comprising one or more processors to perform steps comprising:
receive an object clustering request for one or more objects associated with a plurality of contextual parameters, wherein the plurality of contextual parameters comprises one or more physical attributes and one or more non-physical attributes;
tag the one or more non-physical attributes respectively to the one or more physical attributes;
identify a common context from the one or more physical attributes associated with the one or more objects based on the tagging;
map the one or more physical attributes to the one or more objects based on the common context; and
determine a mean value of the one or more physical attributes by generating a histogram, and a weightage is assigned to the one or more physical attributes;
compute a variance value of the one or more physical attributes from the mean value of the histogram; and
cluster the one or more objects based on the mapping when a distance of the variance value is lesser than a pre-defined distance with respect to the mean value.

Assignees

• Wipro Ltd

Inventors

• Ramachandra Iyer Manjunath
• Pratap Rameshwar
• Chaitanya Villari Boby

Classifications

• G06V10/763

  Non-hierarchical techniques, e.g. based on statistics of modelling distributions · CPC title

• G06V10/776

  Validation; Performance evaluation · CPC title

• G06V40/10

  Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands · CPC title

• G06N20/00 Primary

  Machine learning · CPC title

• G06V10/768

  using context analysis, e.g. recognition aided by known co-occurring patterns · CPC title