Loading...
Please wait, while we are loading the content...
Similar Documents
Kinship detection using facial images
| Content Provider | Indraprastha Institute of Information Technology, Delhi |
|---|---|
| Author | Kohli, Naman |
| Abstract | Face Recognition has improved multiple folds over the years but still remains susceptible to challenges such as illumination, pose and aging. Using context based information is one way to increase the accuracy of a facial recognition system. In this research, we explore kinship detection and kinship as a source of context based information for face recognition. The process of automatically detecting kinship in facial images is a challenging and a relatively unexplored research problem. The reason for this includes less availability of datasets as well as inherent large intra-class variations amongst kins. In this research, IIITD Kinship database was created of over 800+ images pertaining to celebrities ranging over two ethnicities. We first proposed SSD algorithm for detecting kins when a pair of images are given. We also performed research in the cognitive aspect of kinship by conducting a human study to find out the cues that humans use for detecting kins. The observations of human evaluation provided motivation to further enhance the algorithm and propose ’Face Component based Kinship’ algorithm. The results on different databases illustrate the improvement obtained by the proposed algorithm. |
| File Format | |
| Language | English |
| Access Restriction | Authorized |
| Subject Keyword | Kinship Face Recognition DAISY SSD Tri Patch LBP Likelihood Ratio |
| Content Type | Text |
| Educational Degree | Bachelor of Technology (B.Tech.) |
| Resource Type | Thesis |
| Subject | Data processing & computer science |
