Gi Pyo Nam

Intelligent query by humming system based on score level fusion of multiple classifiers

Recently, the necessity for content-based music retrieval that can return results even if a user does not know information such as the title or singer has increased. Query-by-humming (QBH) systems have been introduced to address this need, as they allow the user to simply hum snatches of the tune to find the right song. Even though there have been many studies on QBH, few have combined multiple classifiers based on various fusion methods. Here we propose a new QBH system based on the score level fusion of multiple classifiers. This research is novel in the following three respects: three local classifiers [quantized binary (QB) code-based linear scaling (LS), pitch-based dynamic time warping (DTW), and LS] are employed; local maximum and minimum point-based LS and pitch distribution feature-based LS are used as global classifiers; and the combination of local and global classifiers based on the score level fusion by the PRODUCT rule is used to achieve enhanced matching accuracy. Experimental results with the 2006 MIREX QBSH and 2009 MIR-QBSH corpus databases show that the performance of the proposed method is better than that of single classifier and other fusion methods.

Enhancement of Finger-Vein Image by Vein Line Tracking and Adaptive Gabor Filtering for Finger-Vein Recognition

Biometrics is the technology to identify a user by using the physiological or behavioral characteristics. Among the biometrics such as fingerprint, face, iris, and speaker recognition, finger-vein recognition has been widely used in various applications such as door access control, financial security, and user authentication of personal computer, due to its advantages such as small sized and low cost device, and difficulty of making fake vein image. Generally, a finger-vein system uses near-infrared (NIR) light illuminator and camera to acquire finger-vein images. However, it is difficult to obtain distinctive and clear finger-vein image due to skin scattering of illumination since the finger-vein exists inside of a finger. To solve these problems, we propose a new method of enhancing the quality of finger-vein image. This research is novel in the following three ways compared to previous works. First, the finger-vein lines of an input image are discriminated from the skin area by using local binarization, morphological operation, thinning and line tracing. Second, the direction of vein line is estimated based on the discriminated finger-vein line. And the thickness of finger-vein in an image is also estimated by considering both the discriminated finger-vein line and the corresponding position of finger-vein region in an original image. Third, the distinctiveness of finger-vein region in the original image is enhanced by applying an adaptive Gabor filter optimized to the measured direction and thickness of finger-vein area. Experimental results showed that the distinctiveness and consequent quality of finger-vein image are enhanced compared to that without the proposed method.

Robustness of Face Recognition to Variations of Illumination on Mobile Devices Based on SVM

With the increasing popularity of mobile devices, it has become necessary to protect private information and content in these devices. Face recognition has been favored over conventional passwords or security keys, because it can be easily implemented using a built-in camera, while providing user convenience. However, because mobile devices can be used both indoors and outdoors, there can be many illumination changes, which can reduce the accuracy of face recognition. Therefore, we propose a new face recognition method on a mobile device robust to illumination variations. This research makes the following four original contributions. First, we compared the performance of face recognition with illumination variations on mobile devices for several illumination normalization procedures suitable for mobile devices with low processing power. These include the Retinex filter, histogram equalization and histogram stretching. Second, we compared the performance for global and local methods of face recognition such as PCA (Principal Component Analysis), LNMF (Local Non-negative Matrix Factorization) and LBP (Local Binary Pattern) using an integer-based kernel suitable for mobile devices having low processing power. Third, the characteristics of each method according to the illumination variations are analyzed. Fourth, we use two matching scores for several methods of illumination normalization, Retinex and histogram stretching, which show the best and 2 nd best performances, respectively. These are used as the inputs of an SVM (Support Vector Machine) classifier, which can increase the accuracy of face recognition. Experimental results with two databases (data collected by a mobile device and the AR database) showed that the accuracy of face recognition achieved by the proposed method was superior to that of other methods.

A new query-by-humming system based on the score level fusion of two classifiers

With the widespread use of multimedia devices, such as MP3 players, the necessity of a content-based retrieval is increased, which can find the stored music even if a user does not know the title or singer of the music. Consequently, a query-by-humming (QBH) system is introduced, which provides functionality that a user can find a piece of music by humming. Although there have been many researches into QBH, there has been little done to combine more than two classifiers based on various fusion methods. Hence, we propose a new method of QBH based on the score level fusion of two classifiers. This research is novel in the following three ways as compared with previous works. First, the features of the humming data are extracted by using musical note estimation based on the spectro-temporal autocorrelation (STA). We normalize the extracted features by using the mean-shifting, median filtering, average filtering, and min–max scaling methods. Second, a pitch-based dynamic time warping (DTW) method is used as the first classifier. We use the linear scaling (LS) method with the quantized binary (QB) code of the pitch data as the second classifier. Third, through the combination of these two classifiers based on the score level by the MIN rule, the performance of QBH is much enhanced. Experimental results with the 2006 MIREX QBSH and 2009 MIR-QBSH corpus databases showed that the performance of the proposed fusion method was best compared with single classifier and other fusion methods. Copyright

New Fuzzy-based Retinex Method for the Illumination Normalization of Face Recognition

We propose a new illumination normalization for face recognition which robust in relation to the illumination variations on mobile devices. This research is novel in the following five ways when compared to previous works: (i) a new fuzzy-based Retinex method is proposed for illumination normalization; (ii) the performance of face recognition is enhanced by determining the optimal parameter of Retinex filtering based on fuzzy logic; (iii) the output of the fuzzy membership function is adaptively determined based on the mean and standard deviations of the grey values of the detected face region; (iv) through the comparison of various defuzzification methods in terms of the accuracy of face recognition, one optimal method was selected; (v) we proved the validations of the proposed method by testing it with various face recognition methods. Experimental results showed that the accuracy of the face recognition with the proposed method was enhanced compared to previous ones.

A Study of Deep CNN-Based Classification of Open and Closed Eyes Using a Visible Light Camera Sensor

The necessity for the classification of open and closed eyes is increasing in various fields, including analysis of eye fatigue in 3D TVs, analysis of the psychological states of test subjects, and eye status tracking-based driver drowsiness detection. Previous studies have used various methods to distinguish between open and closed eyes, such as classifiers based on the features obtained from image binarization, edge operators, or texture analysis. However, when it comes to eye images with different lighting conditions and resolutions, it can be difficult to find an optimal threshold for image binarization or optimal filters for edge and texture extraction. In order to address this issue, we propose a method to classify open and closed eye images with different conditions, acquired by a visible light camera, using a deep residual convolutional neural network. After conducting performance analysis on both self-collected and open databases, we have determined that the classification accuracy of the proposed method is superior to that of existing methods.

Intelligent Query by Humming System

This research proposes new system which finds the music by using Query-by-Humming (QBH). For finding a stored music, the features of humming data are selected by using G.729 feature extractor. We normalize the extracted features by using mean-shifting, median filtering, average filtering and min-max scaling methods. Then the corresponding music is matched based on dynamic time warping (DTW) algorithm. As experiment, we compared the matching performance by using the database of Roger Jangs Corpus in MIREX.

Periocular-based biometrics robust to eye rotation based on polar coordinates

Conventional iris recognition requires a high-resolution camera equipped with a zoom lens and a near-infrared illuminator to observe iris patterns. Moreover, with a zoom lens, the viewing angle is small, restricting the user’s head movement. To address these limitations, periocular recognition has recently been studied as biometrics. Because the larger surrounding area of the eye is used instead of iris region, the camera having the high-resolution sensor and zoom lens is not necessary for the periocular recognition. In addition, the image of user’s eye can be captured by using the camera having wide viewing angle, which reduces the constraints to the head movement of user’s head during the image acquisition. Previous periocular recognition methods extract features in Cartesian coordinates sensitive to the rotation (roll) of the eye region caused by in-plane rotation of the head, degrading the matching accuracy. Thus, we propose a novel periocular recognition method that is robust to eye rotation (roll) based on polar coordinates. Experimental results with open database of CASIA-Iris-Distance database (CASIA-IrisV4) show that the proposed method outperformed the others.

Fast query-by-singing/humming system that combines linear scaling and quantized dynamic time warping algorithm

We newly propose a query-by-singing/humming (QbSH) system considering both the preclassification and multiple classifier-based method by combining linear scaling (LS) and quantized dynamic time warping (QDTW) algorithm in order to enhance both the matching accuracy and processing speed. This is appropriate for the QbSH of high speed in the huge distributed server environment. This research is novel in the following three ways. First, the processing speed of the QDTW is generally much slower than the LS method. So, we perform the QDTW matching only in case that the matching distance by LS algorithm is smaller than predetermined threshold, by which the entire processing time is reduced while the matching accuracy is maintained. Second, we use the different measurement method of matching distance in LS algorithm by considering the characteristics of reference database. Third, we combine the calculated distances of LS and QDTW algorithms based on score level fusion in order to enhance the matching accuracy. The experimental results with the 2009 MIR-QbSH corpus and the AFA MIDI 100 databases showed that the proposed method reduced the total searching time of reference data while obtaining the higher accuracy compared to the QDTW.

Periocular Recognition Based on LBP Method and Matching by Bit-Shifting

Periocular recognition requires neither a high-resolution camera nor a zoom lens. It matches using the features extracted from the surrounding area of the eye. In addition, by using a wide-view camera, the constraints to users’ head movement decrease. In this research, we newly propose a periocular recognition based on LBP method and matching by bit-shifting. Our research is novel in the following three manners. First, the iris and pupil region in the input eye image are detected. This allows the accurate eye region to be obtained for periocular recognition. Second, the feature code is extracted from the eye region with a local binary pattern method. Third, the proposed system performs matching by bit-shifting to prevent degradation to the matching accuracy caused by head movement. Experimental results show that the high accuracy of periocular recognition is obtained by the proposed method.