Convolutional neural networks approach for multimodal biometric identification system using the fusion of fingerprint, finger-vein and face images

Abstract

In recent years, the need for security of personal data is becoming progressively important. In this regard, the identification system based on fusion of multibiometric is most recommended for significantly improving and achieving the high performance accuracy. The main purpose of this paper is to propose a hybrid system of combining the effect of tree efficient models: Convolutional neural network (CNN), Softmax and Random forest (RF) classifier based on multi-biometric fingerprint, finger-vein and face identification system. In conventional fingerprint system, image pre-processed is applied to separate the foreground and background region based on K-means and DBSCAN algorithm. Furthermore, the features are extracted using CNNs and dropout approach, after that, the Softmax performs as a recognizer. In conventional fingervein system, the region of interest image contrast enhancement using exposure fusion framework is input into the CNNs model. Moreover, the RF classifier is proposed for classification. In conventional face system, the CNNs architecture and Softmax are required to generate face feature vectors and classify personal recognition. The score provided by these systems is combined for improving Human identification. The proposed algorithm is evaluated on publicly available SDUMLA-HMT real multimodal biometric database using a GPU based implementation. Experimental results on the datasets has shown significant capability for identification biometric system. The proposed work can offer an accurate and efficient matching compared with other system based on unimodal, bimodal, multimodal characteristics.

The classification of EEG-based winking signals: a transfer learning and random forest pipeline

“…The combination of a few decision trees to mitigates the notion between the variance and bias, which in turn reduces the possibility of overfitting. It is worth noting that the RF classifier has been widely used in many different medical oriented types of research ( Cherrat, Alaoui & Bouzahir, 2020 ; Tabares-Soto et al, 2020 ). The RF hyperparameters evaluated in this study are the number of trees (n_estimators), depth of the trees (max_depth), and the measurement of the splitting quality (criterion).…”

Feature Extraction for Finger-Vein-Based Identity Recognition

“…Lastly, the two scores are fused together to provide the final decision. Cherrat et al [ 127 ], in their finger vein system, used a CNN as a feature extractor combined with a Random Forest model for the classification, while Zhao et al [ 128 ] used a lightweight CNN for the classification and focused on the loss function by using the center loss function and dynamic regularization. Hao et al [ 129 ] proposed a multi-tasking neural network that performs both ROI and feature extraction sequentially, through two branches.…”

Comparison of levels and fusion approaches for multimodal biometrics

“…The fusion at the decision is only used where the decisions of the individual users are accessible hence is known as an abstract level fusion [45]. Comparative analysis of of various levels of fusion [46] is shown in Table 2. The noise of sensed data, low sensor efficiency, and atmospheric influences.…”