Open Access

Gender determination from periocular images using deep learning based EfficientNet architecture

Viji B Nambiar

Viji B Nambiar

Department of Mathematics, CHRIST (Deemed to be University)Bengaluru, India
Search for this author on
Sciendo|Google Scholar
Nambiar, Viji B
, 
Bojan Ramamurthy

Bojan Ramamurthy

Department of Computer Science, CHRIST (Deemed to be University)Bengaluru, India
Search for this author on
Sciendo|Google Scholar
Ramamurthy, Bojan
 and 
Pundikala Veeresha

Pundikala Veeresha

Department of Mathematics, CHRIST (Deemed to be University)Bengaluru, India
Search for this author on
Sciendo|Google Scholar
Veeresha, Pundikala
  
Oct 31, 2023
Gender determination from periocular images using deep learning based EfficientNet architecture's Cover Image
About this article
Previous Article
Next Article
Cite
Download Cover
Article Category: Original Study
Published Online: Oct 31, 2023
Page range: 59 - 70
Received: Jul 22, 2023
Accepted: Sep 08, 2023
DOI: https://doi.org/10.2478/ijmce-2024-0005
Keywords
© 2024 Viji B Nambiar et al., published by Sciendo
This work is licensed under the Creative Commons Attribution 4.0 International License.

Fig. 1

Periocular region.
Periocular region.

Fig. 2

CNN architecture.
CNN architecture.

Fig. 3

Stem layer for EfficientNet-B1 [19].
Stem layer for EfficientNet-B1 [19].

Fig. 4

Architecture for EfficientNet-B1 [19].
Architecture for EfficientNet-B1 [19].

Fig. 5

Modules for EfficientNet-B1 [19].
Modules for EfficientNet-B1 [19].

Fig. 6

A subset of the dataset.
A subset of the dataset.

Fig. 7

Confusion matrix for CNN model built from scratch.
Confusion matrix for CNN model built from scratch.

Fig. 8

Model accuracy for CNN built from scratch.
Model accuracy for CNN built from scratch.

Fig. 9

Model loss for CNN built from scratch.
Model loss for CNN built from scratch.

Fig. 10

Confusion matrix for CNN model built from scratch.
Confusion matrix for CNN model built from scratch.

Fig. 11

Model accuracy for EfficientNetB1.
Model accuracy for EfficientNetB1.

Fig. 12

Model loss for EfficientNetB1.
Model loss for EfficientNetB1.

Classification report for fine-tuned EfficientNetB1_

Label Precision Recall F1-score
1 (Male) 0.97 0.99 0.98
0 (Female) 0.99 0.97 0.98

Classification report for CNN model built from scratch_

Label Precision Recall F1-score
1 (Male) 0.94 0.96 0.95
0 (Female) 0.95 0.93 0.94
Language:
English
Publication timeframe:
2 times per year
Journal Subjects:
Computer Sciences, Computer Sciences, other, Engineering, Introductions and Overviews, Engineering, other, Mathematics, General Mathematics, Physics, Physics, other
Journal RSS Feed