An Intelligent Human Age and Gender Forecasting Framework Using Deep Learning Algorithms

Figures & data

Figure 1. Block diagram of age and gender identification approach.

Figure 2. Flowchart of proposed methodology.

Figure 3. SNCNN framework of classification.

Table

Algorithm 1. Spiking Neuron based CNN

Input: Feature vectors from neuro multi-kernel SVM

Step 1: Construction of convolution connection Step 2: Feature vectors are passed via general convolution layer and Semantic information are captured in max pooling layer Step 3: Regularization rate is assigned as 0.4 and features are retrieved from feature vector in drop out layer Step 4: The softmax values are estimated by the evidence of $Evidenc{e}_p=\sum _q{W}_p,q{I}_q+b_p$ Step 5: The probability is detected by $softmax{\left(I\right)}_p=\frac{\mathrm{e}\mathrm{x}\mathrm{p}\left(I_p\right)}{\sum _j\mathrm{e}\mathrm{x}\mathrm{p}\left(I_p\right)}$ Step 6: The weights in the network is optimized with ADAM optimizer Step 7: Relevant information are classified based on the class labels in the output layer

Figure 4. Input image and pre-processed image.

Figure 5. Segmented image – DCNN.

Figure 6. Comparison of age classification.

Table 1. Comparison of age classification

Method	Accuracy (%)	Specificity (%)	Sensitive (%)
Random Forest (RF)	91	86	95.5
Naïve Bayes (NB)	76	79	76
Convolutional Neural Network (CNN)	77.5	80	75
Support Vector Machine (SVM)	83	80.5	85
Deep CNN	91	90	100
SNCNN	99.6	100	100

Figure 7. Comparison of gender classification.

Table 2. Comparison of gender classification

Method	Accuracy (%)	Specificity (%)	Sensitive (%)
Random Forest (RF)	90	85	95
Naïve Bayes (NB)	77.5	80	75
Convolutional Neural Network (CNN)	78	81	76
Support Vector Machine (SVM)	82.5	80	85
Deep CNN	91.7	85.7	100
SNCNN	93.8	100	88.9

Figure 8. ROC curve.

Figure 9. Confusion matrix.

Figure 10. Comparison of loss function and accuracy vs. iteration.

Figure 11. Classification of age and gender.