Method of image-based relationship analysis and system thereof

A computerized system of image-based relationship analysis, the system comprising a processing and memory circuitry (PMC) configured to:

• obtain a set of target images each including one or more image representations of one or more individuals;
• obtain, for each image representation in each target image, a corresponding vector representation generated by a Facial Recognition Model (FRM), thereby providing multiple vector representations corresponding to multiple image representations of individuals included in the set of target images;
• cluster the multiple vector representations to a plurality of clusters of vector representations corresponding to a plurality of unique individuals included in the set of target images using a similarity measure, and obtaining, for each target image, one or more unique individuals associated therewith;
• for each given target image of at least one subset of the set,
• obtain a set of image parameters characterizing the given target image, wherein the set of image parameters includes at least one computed parameter indicative of a relationship measurement between the one or more unique individuals associated with the given target image;
• generate a local relationship matrix using the set of image parameters, wherein the local relationship matrix is representative of local mutual relationships between the one or more unique individuals;
• thereby obtaining a set of local relationship matrices corresponding to the at least one subset of target images;
• and
• generate a global relationship matrix by combining the set of local relationship matrices, the global relationship matrix being representative of relationships between the plurality of unique individuals.

A non-transitory computer readable medium comprising instructions that, when executed by a computer, cause the computer to perform a method of image-based relationship analysis, the method comprising:

• obtaining a set of target images each including one or more image representations of one or more individuals;
• obtaining, for each image representation in each target image, a corresponding vector representation generated by a Facial Recognition Model (FRM), thereby providing multiple vector representations corresponding to multiple image representations of individuals included in the set of target images;
• clustering the multiple vector representations to a plurality of clusters of vector representations corresponding to a plurality of unique individuals included in the set of target images using a similarity measure, and obtaining, for each target image, one or more unique individuals associated therewith;
• for each given target image of at least one subset of the set,
• obtaining a set of image parameters characterizing the given target image, wherein the set of image parameters includes at least one computed parameter indicative of a relationship measurement between the one or more unique individuals associated with the given target image;
• generating a local relationship matrix using the set of image parameters, wherein the local relationship matrix is representative of local mutual relationships between the one or more unique individuals;
• thereby obtaining a set of local relationship matrices corresponding to the at least one subset of target images;
• and
• generating a global relationship matrix by combining the set of local relationship matrices, the global relationship matrix being representative of relationships between the plurality of unique individuals.