System and method for detecting similarities
Systems and methods for determining similarities between an input data set and a target data set with the data sets being vector representations of the features of a candidate potential copy and a target original. A feature extraction module receives an image of the potential copy and extracts the features of that candidate. The features of the target original may already be extracted or may be separately extracted. The resulting data sets for the candidate and the original are then passed through a decision module. The decision module determines a level of similarity between the features of the candidate and the features of the original. The output of the decision module provides an indication of this level of similarity and, depending on this level of similarity, an alert may be generated. A report module may be included to provide an explanation regarding the level of similarity.
1. A computer-implemented method for determining a similarity between an input data set detailing features relating to an input product signature and a target data set detailing features relating to a target, the computer-implemented method being executed by at least one processor operatively connected to a non-transitory storage medium, the computer-implemented method comprising:
extracting a vector representation of features of said input product signature to result in said input data set;
processing said input data set to determine a level of similarity between said features of said input product signature and said features of said target, and to detect and assess similarities between said features of said target in said target data set and features in said input product signature;
receiving a data output from said processing step, said data output being indicative of a level of similarity between said features detailed in said input data set and said features detailed in said target data set;
in the event said data output conforms to at least one predetermined criterion, generating an alert;
sending said data output to a user for validation; and
receiving results of said validation;
wherein said data output and said results of said validation are further processed for use in further training a model used in determining said level of similarity only when results of said validation and said data output indicate an error in said level of similarity.
2. The method according to claim 1 , wherein said input product signature contains at least one of:
a trademark;
a logo;
business related indicial;
marketing material relating to a specific product;
product packaging relating to a specific product; and
at least one image of a specific product; and
wherein said target data set detailing features relating to said target is derived by an analysis of at least one of:
a registered industrial design for a target product;
product packaging for a target product;
a copyrighted image;
a trademark;
a logo;
business related indicia; and
text relating to a target product.
3. The method according to claim 1 , wherein said at least one predetermined criterion comprises said data output indicating a level of similarity above a predetermined level.
4. The method according to claim 1 , wherein said target data set is derived from one or more written descriptions of said target.
5. The method according to claim 1 , further comprising a step of generating an explanation heat map outlining aspects of said input product signature and said target to thereby provide an explanation of said data output.
6. A system for detecting similarities between at least one candidate product signature and a target, the system comprising: non-transitory storage medium storing instructions thereon; and at least one processor operatively connected to the non-transitory storage medium, the processor, upon executing the instructions, being configured for:
receiving said at least one candidate product signature and producing a vector representation of features of said at least one candidate product signature;
processing said vector representation of features of said at least one candidate product signature and comparing features of said at least one candidate product signature and features of said target;
wherein
an output of said processing is indicative of a level of similarity between features in said at least one candidate product signature and features of said target, and
the system is further configured for sending said output to a user for validation and configured for receiving user supplied validation results as training inputs to produce better similarity results.
7. The system according to claim 6 , wherein said processing produces a similarity score detailing said level of similarity between said features in said candidate product signature and said features of said target.
8. The system according to claim 7 , further configured for generating a report for a user whenever said similarity score exceeds a predetermined threshold, said report providing details regarding said similarity score.
9. The system according to claim 8 , wherein said report also includes explanatory images.
10. The system according to claim 6 , wherein only negative validation results are used to produce better similarity results.
11. The system according to claim 6 , wherein said system generates an alert when an output of said processing conforms to at least one predetermined criterion.
12. The system according to claim 11 , wherein said at least one predetermined criterion comprises said output of said processing indicating a level of similarity above a predetermined level.
13. The system according to claim 6 , wherein said system outputs an explanation heat map that provides an explanation regarding said output.
14. The system according to claim 13 , wherein said system is used for quality assurance purposes.
15. The system according to claim 14 , wherein said system is used to detect defects in manufactured products by comparing a defect-free sample product and an end product.
16. The system according to claim 6 , wherein said receiving said at least one candidate product signature and said producing said vector representation further includes a neural network trained to extract features of images and data associated with a candidate.
17. A system for detecting similarities between at least one candidate product signature and a target, the system comprising: non-transitory storage medium storing instructions thereon; and at least one processor operatively connected to the non-transitory storage medium, the processor, upon executing the instructions, being configured for:
receiving said at least one candidate product signature and producing a vector representation of features of said at least one candidate product signature;
processing said vector representation of features of said at least one candidate product signature and comparing features of said at least one candidate product signature and features of said target;
wherein
an output of said processing is indicative of a level of similarity between features in said at least one candidate product signature and features of said target,
said receiving said at least one candidate product signature and said producing said vector representation further includes a neural network trained to extract features of images and data associated with a candidate, and
text associated with the candidate is processed by a separate text feature extractor.
18. The system according to claim 17 , output representations form a joint representation including image and text.