Analysis of facial imagery
Employer
[no-member:pro]Igor[/no-member:pro]
Project parameters
Type of cooperationOne-time project
SectionSoftware development
Prepaymentwithout prepayment
Payment methodsBank transfer
Acceptance of requestsfrom Aug 14, 2022 until Aug 19, 2022
Project description
Programmer services are required. We need to analyze facial images. Two groups of images of men and women are provided - a test group and a control group.
The images in the test group are divided into pairs - a male woman (married). The images in the control group are random images of men and women.
We are going to use several different comparison algorithms – the standard one used to identify photos of the same person (confidence level and will be the similarity level), as well as our own algorithm. The relevant instruction will be provided.
The result of the work should be:
Two databases (two groups) of “recognized” faces – that is, the coordinates of all recognizable points of the face.
List of all married couples and measure of the “similarity” of spouses according to the classical algorithm of comparison.
Average “similarity” among couples and average “similarity” among the control sample.
Software code for continuing research.
Before the recognition begins, “bad”, blurred, faceless photos, etc. Must be removed.
The images in the test group are divided into pairs - a male woman (married). The images in the control group are random images of men and women.
We are going to use several different comparison algorithms – the standard one used to identify photos of the same person (confidence level and will be the similarity level), as well as our own algorithm. The relevant instruction will be provided.
The result of the work should be:
Two databases (two groups) of “recognized” faces – that is, the coordinates of all recognizable points of the face.
List of all married couples and measure of the “similarity” of spouses according to the classical algorithm of comparison.
Average “similarity” among couples and average “similarity” among the control sample.
Software code for continuing research.
Before the recognition begins, “bad”, blurred, faceless photos, etc. Must be removed.
