UFPR Preliminary Results
Dataset
The UFPR-Periocular dataset was created to obtain images in unconstrained scenarios that contain realistic noises caused by occlusion, blur, and variations in lighting, distance, and angles.
The gender distribution of the subjects is $(53,65\%)$ male and $(46,35\%)$ female, and approximately $66\%$ of the subjects are under $31$ years old. In total, the dataset has images captured from $196$ different mobile devices – the five most used device models were: Apple iPhone 8 $(4.1\%)$, Apple iPhone 9 $(3.1\%)$, Xiaomi Mi 8 Lite $(3.0\%)$, Apple iPhone 7 $(3.0\%)$, and Samsung Galaxy J7 Prime $(2.7\%)$.
Gender Classification Results
Subject Verification Results
Subject Verification All Results Table
Visualizations
GradCAM
GradCAM
Guided GradCAM
Guided GradCAM
Occlusion Sensitivity
Occlusion Sensitivity
Vanilla Gradients
Vanilla Gradients
Integrated Gradients
Integrated Gradients
SmoothGrad
SmoothGrad
Gradients x Input
Gradients x Input
References
Bibliography called, but no references
Sreeraj Ramachandran
Graduate Research Assistant
My research interests include Computer Vision, Biometrics, Bias AI, GANs and adversarial Attacks