Age and gender estimation using deep residual learning network

Lee, Seok Hee; Hosseini, Sepidehsadat; kwon, hyuk jin; Moon, Jaewon; Koo, Hyung Il; Cho, Nam Ik

doi:10.1109/iwait.2018.8369763

Cited by 19 publications

(12 citation statements)

References 5 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The identity connections add neither extra parameters nor computational complexity to the network. Residual networks have been used successfully in age and gender estimation [ 28 ], for hyperspectral image classification [ 29 ], and other classification tasks. Even though very deep residual networks (152 layers) have been used, in this work, we only used 50 layers, as our focus was on a comparative evaluation of architectures.…”

Section: Proposed Methodsmentioning

confidence: 99%

Optimizing Deep CNN Architectures for Face Liveness Detection

Koshy

Mahmood

2019

Entropy

View full text Add to dashboard Cite

Face recognition is a popular and efficient form of biometric authentication used in many software applications. One drawback of this technique is that it is prone to face spoofing attacks, where an impostor can gain access to the system by presenting a photograph of a valid user to the sensor. Thus, face liveness detection is a necessary step before granting authentication to the user. In this paper, we have developed deep architectures for face liveness detection that use a combination of texture analysis and a convolutional neural network (CNN) to classify the captured image as real or fake. Our development greatly improved upon a recent approach that applies nonlinear diffusion based on an additive operator splitting scheme and a tridiagonal matrix block-solver algorithm to the image, which enhances the edges and surface texture in the real image. We then fed the diffused image to a deep CNN to identify the complex and deep features for classification. We obtained 100% accuracy on the NUAA Photograph Impostor dataset for face liveness detection using one of our enhanced architectures. Further, we gained insight into the enhancement of the face liveness detection architecture by evaluating three different deep architectures, which included deep CNN, residual network, and the inception network version 4. We evaluated the performance of each of these architectures on the NUAA dataset and present here the experimental results showing under what conditions an architecture would be better suited for face liveness detection. While the residual network gave us competitive results, the inception network version 4 produced the optimal accuracy of 100% in liveness detection (with nonlinear anisotropic diffused images with a smoothness parameter of 15). Our approach outperformed all current state-of-the-art methods.

show abstract

Section: Proposed Methodsmentioning

confidence: 99%

Optimizing Deep CNN Architectures for Face Liveness Detection

Koshy

Mahmood

2019

Entropy

View full text Add to dashboard Cite

show abstract

“…To further verify the validity of the method, the results are compared with other age estimation methods based on deep learning, and the results are shown in TABLE II and Figure 8 and 9. [6] 4.67 / Levi [7] 3.85 / Lee [8] 3.80 3.74 Li [9] 3.73 / Singhal [10] 3.44 / Deep embedding method [49] 3.32 3.71 TF [50] 3.277 3.35 Apparent methods [51] 3.272 / Cluster-CNN [52] 3.24 3.43 LAAE 3.04(d=8) 3.17(d=4)…”

Section: Contrast Experimentsmentioning

confidence: 99%

“…After realizing the probability predictions of different age groups, their aggregate them by converting them to calculate the value distribution of the entire age group, and let them get the final age estimate from their votes. Lee [8] proposed a deep residual learning model for age and gender estimation. Their method detects faces in the input image, and then estimates the age and gender of each face.…”

Section: Introductionmentioning

confidence: 99%

Combined Deep Learning With Directed Acyclic Graph SVM for Local Adjustment of Age Estimation

et al. 2021

View full text Add to dashboard Cite

“…where y' and y present the predicted and real age value respectively and N denotes the number of the testing facial images. The purpose from our work is not to extract exactly the age but we look to just classify ages into three ranges, youth (16)(17)(18)(19)(20)(21)(22)(23)(24)(25)(26)(27)(28)(29)(30), senior (31-50) and elderly (51-over). The proposed method obtains an MAE of 3.26 years, which is considerable very important compared with other methods.…”

Section: Age Gender and Ethnicity Recognition (Ager)mentioning

confidence: 99%

“…Table 3. Comparison of gender accuracy with the state-of-the-art methods (%) Method Year Accuracy Duan et al [8] 2017 88.20 Guo et al [10] 2014 98.40 Dhomne et al [16] 2018 95.00 Srinivas et al [17] 2017 84.70 Lee et al [20] 2017 88.50 Huang et al [21] 2017 89.60 Benini et al [22] 2019 98.59 Fang et al [23] 2019 98.80 Proposed method -95.00 Table 4. Accuracy of ethnicity recognition for MORPH II dataset (%) Method Year Accuracy Guo et al [10] 2014 99.00 Uddin et al [13] 2016 95.40 Srinivas et al [17] 2017 33.33 Mohammed et al [18] 2019 93.3 Hocquet et al [24] 2016 97.50 Mohammed et al [25] 2017 94.60 Proposed method -98.20…”

Section: Age Gender and Ethnicity Recognition (Ager)mentioning

confidence: 99%

Semi-supervised auto-encoder for facial attributes recognition

2020

View full text Add to dashboard Cite

The particularity of our faces encourages many researchers to exploit their features in different domains such as user identification, behaviour analysis, computer technology, security, and psychology. In this paper, we present a method for facial attributes analysis. The work addressed to analyse facial images and extract features in the purpose to recognize demographic attributes: age, gender, and ethnicity (AGE). In this work, we exploited the robustness of deep learning (DL) using an updating version of autoencoders called the deep sparse autoencoder (DSAE). In this work we used a new architecture of DSAE by adding the supervision to the classic model and we control the overfitting problem by regularizing the model. The pass from DSAE to the semi-supervised autoencoder (DSSAE) facilitates the supervision process and achieves an excellent performance to extract features. In this work we focused to estimate AGE jointly. The experiment results show that DSSAE is created to recognize facial features with high precision. The whole system achieves good performance and important rates in AGE using the MORPH II database

show abstract

Age and gender estimation using deep residual learning network

Cited by 19 publications

References 5 publications

Optimizing Deep CNN Architectures for Face Liveness Detection

Optimizing Deep CNN Architectures for Face Liveness Detection

Combined Deep Learning With Directed Acyclic Graph SVM for Local Adjustment of Age Estimation

Semi-supervised auto-encoder for facial attributes recognition

Contact Info

Product

Resources

About