Enhanced Face Presentation Attack Prevention Employing Feature Fusion of Pre-trained Deep Convolutional Neural Network Model and Thepade's Sorted Block Truncation Coding

Thepade, S. D.; Dindorkar, M. R.; Chaudhari, Pradip; Bang, S. V.

doi:10.5829/ije.2023.36.04a.17

Cited by 2 publications

(3 citation statements)

References 36 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In these networks, a large number of layers in these networks are trained with a powerful approach. This technique is extremely effective and is also frequently utilized in a variety of computer vision applications [3,4]. The major CNN was AlexNet, introduced by Krizhevsk [5] fourteen years after LeNet in 2017.…”

Section: Deep Neural Networkmentioning

confidence: 99%

See 1 more Smart Citation

Deep Multi-task Convolutional Neural Networks for Efficient Classification of Face Attributes

Rohani,

Farsi,

Mohamadzadeh

2023

IJE

View full text Add to dashboard Cite

Facial feature recognition is an important subject in computer vision with numerous applications. The human face plays a significant role in social interaction and personology. Valuable information such as identity, age, gender, and emotions can be revealed via facial features. The purpose of this paper is to present a technique for detecting age, smile, and gender from facial images. A multi-task deep learning (MT-DL) framework was proposed that can simultaneously estimate three important features of the human face with remarkable accuracy. Additionally, the proposed approach aims to reduce the number of trainable network parameters while leveraging the combination of features from different layers to increase the overall accuracy. The conducted tests demonstrate that the proposed method outperforms recent advanced techniques in all three accuracy criteria. Moreover, it was demonstrated that multi-task learning (MTL) is capable of improving the accuracy by 1.55% in the smile task, 2.04% in the gender task, and 3.52% in the age task even with less available data, by utilizing tasks with more available data. Furthermore, the trainable parameters of the network in the MTL mode for estimating three tasks simultaneously increase only by about 40% compared to the single-task mode. The proposed method was evaluated on the IMDB-WIKI and GENKI-4K datasets and produced comparable accuracy to the state-of-the-art methods in terms of smile, age detection, and gender classification.

show abstract

Section: Deep Neural Networkmentioning

confidence: 99%

“…An exponential function was used to reduce the learning rate. This rate is calculated in the p th step of Equation (4).…”

Section: Simulation Detailsmentioning

confidence: 99%

Deep Multi-task Convolutional Neural Networks for Efficient Classification of Face Attributes

Rohani,

Farsi,

Mohamadzadeh

2023

IJE

View full text Add to dashboard Cite

show abstract

“…Numerical and analytical models have been widely used to analyze the fundamental behavior of mechanical systems [21][22][23]. For example, Abdollahi et al [24] developed a finite element model (FEM) to analyze the impact of hybrid nanofluid on the flow field and heat transfer in parallel surfaces.…”

Section: Introductionmentioning

confidence: 99%

Application of Ultra-sensitive Pillar-enhanced Quartz Crystal Resonators for Airborne Detection of Nanoparticles: A Theoretical Study

Xie

Zheng

2023

IJE

View full text Add to dashboard Cite

Although quartz crystal resonators (QCR) have been used for airborne detection of particles and viruses, they suffer from various limitations, such as low sensitivity compared to other devices. Therefore, it is necessary to develop a new device capable of achieving high sensitivity, which can be used for practical airborne detections. The current study reports a comprehensive parametric theoretical model for analyzing the response of ultra-sensitive pillar-enhanced QCR (QCR-P) for airborne detection of nanoparticles. The electromechanical model comprised an equivalent circuit integrated with pillars containing nanoparticles. It was shown that pillar height and particle radius play a critical role in the response of QCR-P devices. The study revealed that selecting the optimal pillar height can lead to a significant frequency shift depending on the nanoparticle radius and pillar height, while it is independent of particle mass density. These results underscore the potential of utilizing pillars to substantially enhance the sensitivity of conventional QCR up to 140 times in the airborne detection of nanoparticles. These findings can be utilized to design optimum pillar heights to achieve maximum sensitivity in the airborne detection of nanoparticles and proteins, thereby enabling the adoption of ultra-sensitive pillarenhanced quartz crystal resonators for practical airborne applications.

show abstract

Enhanced Face Presentation Attack Prevention Employing Feature Fusion of Pre-trained Deep Convolutional Neural Network Model and Thepade's Sorted Block Truncation Coding

Cited by 2 publications

References 36 publications

Deep Multi-task Convolutional Neural Networks for Efficient Classification of Face Attributes

Deep Multi-task Convolutional Neural Networks for Efficient Classification of Face Attributes

Application of Ultra-sensitive Pillar-enhanced Quartz Crystal Resonators for Airborne Detection of Nanoparticles: A Theoretical Study

Contact Info

Product

Resources

About