![]() Section 4 describes the overall structure of the proposed method. In Section 3, the selected feature extractors to be used in feature-level and score-level fusion are reviewed. The rest of the paper is organized as follows: previous studies in facial age estimation have been reviewed in Section 2. Additionally, for the first time, we show that the feature-level and score-level fusion of textural descriptors, local appearance-based feature extractors, and global feature descriptors achieve superior result than the other previous works in this field. The main contributions are as follows: a brief summary of several efficient local and global feature extractors for image classification which are used for the first time in age estimation problem is given and investigated their discrimination power. In this study, we evaluate different popular local and global feature extractors in both feature-level and score-level fusion in order to investigate their efficiency and discrimination in the age estimation field. Patil and Bhalke ( 2016) proposed a multibiometric system that used three traits such as fingerprint, palmprint, and iris that are combined by using weighted fusion technique for person identification. Marcialis and Roli ( 2007) suggested a score-level fusion of fingerprint and face matchers for personal verification under stress conditions. Score-level fusion has been used in many research studies. Additionally, it is relatively easy to access and combine the scores generated by different biometric matchers. ![]() Apart from the raw data and feature vectors, the match scores contain the richest information about the input pattern. When match scores’ outputs by different biometric matchers are consolidated in order to arrive at a final recognition decision, fusion is said to be done at the match score level which is also known as score-level fusion. ![]() In a biometric recognition system, the match score is a measure of similarity between the input and template biometric feature vectors. 2014 Youssef, Elberrichi, and Adjoudj 2010). Score level fusion shows very good performance in multimodal biometric systems (Marcialis and Roli 2007 Patil and Bhalke 2016 Sim et al. These schemes that determine intensity variations in small area or local neighborhood templates from spatial patches are fundamental approaches for a variety of applications which deal with extremely uncontrolled environments. Recent advances in image classification and object recognition fields helped researchers to propose several very efficient and histogram-based feature extraction approaches that have interesting properties such as invariant to scale and rotation and robust to illumination and alignment variances. In both cases, the generated feature extractor method is difficult to reuse, and in many cases, it has high-dimension which takes considerable time to be extracted. The inherent difficulties in age estimation from facial image, have derived research into constructing especially complex feature descriptor approaches in which most of them are either user-defined multi-level and orientation bank of filters which were tried to mimic the behavior of animal visual cortex (primary) network, or fine-grained facial regions to perform accurate alignment by using multiple facial fiducial points. Experiments on the publicly available MORPH and FG-NET databases prove the effectiveness of the proposed method and the proposed method outperforms many of the state-of-the-art systems. Feature-level fusion of biologically inspired and texture-based methods is integrated into the proposed method and their combination is fused with an appearance-based method using score-level fusion. In our proposed method, the advantage of using different types of features such as biologically inspired features, texture-based features, and appearance-based features is used. This integration is performed by using two-level fusion of features and scores with the help of feature-level and score-level fusion techniques. In this paper, an integration of different type of feature extraction algorithms is applied on facial images for accurate age estimation. Facial feature extraction algorithms play an important role in many applications of face biometrics such as face recognition for person identification, classification of emotions by facial expression recognition and age estimation using facial images.
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |