GENDER AND AGE RECOGNITION FOR VIDEO ANALYTICS SOLUTION PRESENTED BY: SUBHASH REDDY JOLAPURAM

INTRODUCTION INTRODUCTION INTRODUCTION: 1.For Gender and Age recognition an application is used as a video data analysis. 2.Automatic video data analysis is a challenging problem to find a particular object in a video stream and automatically it decides if it belongs to particular class one. 3.For this video data analysis a number of different machine learning techniques and algorithms is used. 4.It is based on computer vision and machine learning methods. \\\\

The Application of Video Data Analysis:

Gender Recognition : It is based on non-linear support vector machine(SVM) classifier with radial basis function(RBF) kernel. The Extract information from image fragment and move to a lower dimension feature space local binary patterns(LBF) is utilized. These features have been proved to show good results in application to face recognition tasks.

Calculation procedure for LBP:

Features of Gender Recognition: In the gender recognition algorithm training and testing require big enough color image database. The most commonly used image database for the tasks of human face recognition is the FERET database. But in this we collected our own image database gathered from different sources. Faces on the images from the proposed database were detected automatically by Adaboost face detection algorithm.

The Image Database Parameters parametervalue The total number of images8654 The number of male faces5250 The number of female faces5250 Minimum image resolution640*480 Color space formatRGB Face posistionFrontal People’s ageFrom 18 to 65 years old RaceCaucasian Lightning conditions, background and facial Expression No restrictions

Comparison of LBP-SVM and AF-SVM For the representation of classification results are utilized the Received operator characteristics (ROC curve). There are two classes one of them is considered to be positive decision and other is negative. Roc curve is created by plotting the fraction of true positives out of the positives(TPR=true positive rate vs. the function of false positives out of the negative (FPR=false positive rate) at various discrimination threshold settings. The Experimental results show that utilization of LBP features for gender recognition improves overall performance by 1.5% and the accuracy is 92%.

Recognition rate for LBP-SVM and AF-SVM AlgorithmAF-SVMLBP-SVM Recognition RateTrue False Classified male% Classified female% Total classification rate%

ROC Curves for LBP-SVM and AF-SVM Classifiers.

AGE Estimation Algorithm The Age Estimation Algorithm realizes multiclass classification approach Image normalization was performed by histogram equalization procedure To predict direct age binary classifier outputs are statistically analyzed and the most probable age becomes the algorithm output. The Test age estimation algorithms are Mean Absolute Error(MAE) and Cummulative Score(CS). There are conventional databases that are widely used in a field of age estimation for facial images.

LBP-SVM Age Estimation Block diagram

MAE and CS of FG-NET and LBP-SVM Algorithm To Estimate the proposed algorithm in real-life situation testing firstly performance on FG-NET database The proposed algorithm shows results comparable to the subjective evaluation in range of ages from 20 to 35 years Analysis of the error probability density function shows that the proposed algorithm has close to systematic error distribution. The Examples of face images where LBP-SVM algorithm have good and poor age estimation.

MAE on FG-NET database for LBP-SVM algorithm

CS on FG-NET database for LBP-SVM algorithm

CONCLUSION A modern efficient machine learning algorithm based on LBP features allows us to recognize viewers gender for video analytics systems with more than 92% accuracy The obtained results show that our framework provides high accuracy for both gender and age estimation classification problems compared to the best known methods. The gender and age estimation algorithms described in this paper has integrated in audience measurement system which can collect and process the video data in real time.