Human iris feature extraction under pupil size variation using local texture descriptors

Souza, Jones Mendonça de; Gonzaga, Adilson

doi:10.1007/s11042-019-7371-4

Cited by 4 publications

(3 citation statements)

References 22 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Comparing the best results obtained using the CASIA-Iris-V4 (Table. VI) in [11] with the ones obtained in this research using the same dataset (Table . I,Table . II,Table. III) we can see a higher success rate in the methodology used in this research, being 95.42% of success rate using the InceptionV1 network with the OneV-ersusAll training algorithm (Table.…”

Section: A Discussionmentioning

confidence: 54%

“…Reference [11] presents iris at a distance recognition, a different methodology capable of performing the task of person recognition proposed in this research. The CASIA-Iris-V4 and UBIRIS.V2 are among the number of datasets utilized in [11], allowing a comparison between both methods of recognition. The results obtained in [11] using the CASIA-Iris-V4 (Table . VI) dataset are superior to the ones obtained with the UBIRIS.V2 dataset (Table.…”

Section: A Discussionmentioning

confidence: 99%

“…The CASIA-Iris-V4 and UBIRIS.V2 are among the number of datasets utilized in [11], allowing a comparison between both methods of recognition. The results obtained in [11] using the CASIA-Iris-V4 (Table . VI) dataset are superior to the ones obtained with the UBIRIS.V2 dataset (Table. VII), similar to the results produced in this research.…”

Section: A Discussionmentioning

confidence: 99%

See 2 more Smart Citations

Deep Learning Utilized for Person Recognition Based on the Biometric Features of the Periocular Region

Nascimeno¹,

Souza²

2022

Anais Estendidos Do XXXV Conference on Graphics, Patterns and Images (SIBGRAPI Estendido 2022)

View full text Add to dashboard Cite

This article proposes the use of deep learning technologies to perform visual biometric recognition. The results obtained by convolutional neural networks trained to perform multi-class classification based on the visual features of the human periocular region are presented and discussed, in addition to being compared with results obtained using pattern recognition for biometric recognition from human iris textures.

show abstract

Section: A Discussionmentioning

confidence: 54%