Sunjin Yu

3D hand tracking using Kalman filter in depth space

Hand gestures are an important type of natural language used in many research areas such as human-computer interaction and computer vision. Hand gestures recognition requires the prior determination of the hand position through detection and tracking. One of the most efficient strategies for hand tracking is to use 2D visual information such as color and shape. However, visual-sensor-based hand tracking methods are very sensitive when tracking is performed under variable light conditions. Also, as hand movements are made in 3D space, the recognition performance of hand gestures using 2D information is inherently limited. In this article, we propose a novel real-time 3D hand tracking method in depth space using a 3D depth sensor and employing Kalman filter. We detect hand candidates using motion clusters and predefined wave motion, and track hand locations using Kalman filter. To verify the effectiveness of the proposed method, we compare the performance of the proposed method with the visual-based method. Experimental results show that the performance of the proposed method out performs visual-based method.

Face liveness detection using variable focusing

As Face Recognition(FR) technology becomes more mature and commercially available in the market, many different anti-spoofing techniques have been recently developed to enhance the security, reliability, and effectiveness of FR systems. As a part of anti-spoofing techniques, face liveness detection plays an important role to make FR systems be more secured from various attacks. In this paper, we propose a novel method for face liveness detection by using focus, which is one of camera functions. In order to identify fake faces (e.g. 2D pictures), our approach utilizes the variation of pixel values by focusing between two images sequentially taken in different focuses. The experimental result shows that our focus-based approach is a new method that can significantly increase the level of difficulty of spoof attacks, which is a way to improve the security of FR systems. The performance is evaluated and the proposed method achieves 100% fake detection in a given DoF(Depth of Field).

Biometric Image Authentication using Watermarking

In this paper, we propose an invertible authentication watermarking algorithm which can detect block-wise malicious manipulations in biometric images. Our method uses an invertible watermark that can also detect manipulated area simultaneously. Virtually all watermarking schemes introduce a small amount of irrecoverable distortion in original biometric images. But our new method is invertible in the sense that, if the data is deemed authentic, distortion due to authentication can be removed if it becomes necessary to obtain the original biometric image. In our method two watermarks are embedded into biometric image. The first one is based on the conventional method which can completely remove the distortion due to authentication if the data is deemed authentic. The second one can detect the block-wise malicious manipulation using the cyclic redundancy check (CRC) concept in the image block. Our proposed method can classify a test image into two types; authentic, and manipulated. Also, this technique provides cryptographic strength when verifying image integrity because the probability of making an undetectable modification to the image can be directly related to a secure cryptographic element, such as a hash function

Robust Face Recognition by Fusion of Visual and Infrared Cues

This paper proposes a robust face recognition method by fusing images acquired from visual and infrared (IR) sensors. Although current 2D image face recognition systems have reached a certain level of maturity, the performance of these systems has been limited by external conditions such as pose, expression and illumination. To alleviate some of these problems, infrared sensor based methods have been suggested. These methods showed very good performance when used in illumination variation environments. However, one of the main drawbacks of using IR sensors for recognition is that they are very sensitive to ambient temperatures. To solve this problem, we suggest a face recognition method that fuses both visual and IR techniques. We developed a visual and IR face image database with photographs taken under a wide range of imaging conditions. With this database, we are able to produce full combinations of comparative experiments in the field of face recognition. Our experiments showed that when using both modalities, results are far better in terms of recognition performance than when using only one modality. The overall average performance was observed to have improved under all imaging conditions

3D face modeling using the multi-deformable method

In this paper, we focus on the problem of the accuracy performance of 3D face modeling techniques using corresponding features in multiple views, which is quite sensitive to feature extraction errors. To solve the problem, we adopt a statistical model-based 3D face modeling approach in a mirror system consisting of two mirrors and a camera. The overall procedure of our 3D facial modeling method has two primary steps: 3D facial shape estimation using a multiple 3D face deformable model and texture mapping using seamless cloning that is a type of gradient-domain blending. To evaluate our methods performance, we generate 3D faces of 30 individuals and then carry out two tests: accuracy test and robustness test. Our method shows not only highly accurate 3D face shape results when compared with the ground truth, but also robustness to feature extraction errors. Moreover, 3D face rendering results intuitively show that our method is more robust to feature extraction errors than other 3D face modeling methods. An additional contribution of our method is that a wide range of face textures can be acquired by the mirror system. By using this texture map, we generate realistic 3D face for individuals at the end of the paper.

An adaptive local binary pattern for 3D hand tracking

Abstract Ever since the availability of real-time three-dimensional (3D) data acquisition sensors such as time-of-flight and Kinect depth sensor, the performance of gesture recognition can be largely enhanced. However, since conventional two-dimensional (2D) image based feature extraction methods such as local binary pattern (LBP) generally use texture information, they cannot be applied to depth or range image which does not contain texture information. In this paper, we propose an adaptive local binary pattern (ALBP) for effective depth images based applications. Contrasting to the conventional LBP which is only rotation invariant, the proposed ALBP is invariant to both rotation and the depth distance in range images. Using ALBP, we can extract object features without using texture or color information. We further apply the proposed ALBP for hand tracking using depth images to show its effectiveness and its usefulness. Our experimental results validate the proposal.

3D Multi-Spectrum Sensor System with Face Recognition

This paper presents a novel three-dimensional (3D) multi-spectrum sensor system, which combines a 3D depth sensor and multiple optical sensors for different wavelengths. Various image sensors, such as visible, infrared (IR) and 3D sensors, have been introduced into the commercial market. Since each sensor has its own advantages under various environmental conditions, the performance of an application depends highly on selecting the correct sensor or combination of sensors. In this paper, a sensor system, which we will refer to as a 3D multi-spectrum sensor system, which comprises three types of sensors, visible, thermal-IR and time-of-flight (ToF), is proposed. Since the proposed system integrates information from each sensor into one calibrated framework, the optimal sensor combination for an application can be easily selected, taking into account all combinations of sensors information. To demonstrate the effectiveness of the proposed system, a face recognition system with light and pose variation is designed. With the proposed sensor system, the optimal sensor combination, which provides new effectively fused features for a face recognition system, is obtained.

Nonintrusive 3-D face data acquisition system

This paper describes a nonintrusive three-dimensional (3-D) face data acquisition system consisting of a stereo vision system and an 850 nm near-infrared line laser. Although a two-dimensional (2-D) face recognition system can achieve a reliable recognition rate, its performance can be degraded by illumination and pose variation. To alleviate these factors in 2-D face recognition, 3-D face recognition has received much attention. To develop a reliable 3-D face recognition system, many researchers have also focused on 3-D face data acquisition. Earlier 3-D face acquisition systems use visible patterns as features to obtain accurate 3-D data, which makes anyone who wants to be verified uncomfortable. In this paper, we propose a novel 850 nm infrared line laser pattern which is almost invisible for 3-D face data acquisition. The reconstructed 3-D face data consists of over 20,000 3-D points; these data can be used effectively for 3-D face recognition.

Iterative three-dimensional head pose estimation using a face normal vector

The performance of face recognition systems has much been burdened by head pose variation. To solve this problem, 3-D face recognition systems that make use of multiple views and depth information have been suggested. However, without an accurate head pose estimation, the performance improvement of 3-D face recognition systems under pose variations remains limited. Previous research on 3-D head pose estimation has been conducted in 3-D space, where the estimation complexity is high. Also it is difficult to incorporate those salient 2-D face features for effective estimation. We propose a novel iterative 3-D head pose estimation method incorporating both 2-D and 3-D face information. To verify the effectiveness, we apply the proposed method to 3-D face modeling and recognition systems with adaptation to various 3-D face data acquisition devices. Our experimental results show that the proposed method can be very effective in terms of modeling and recognition applications, particularly on combining different kinds of acquisition devices, which use different coordinates of origin and scale.

Random-Profiles-Based 3D Face Recognition System

In this paper, a noble nonintrusive three-dimensional (3D) face modeling system for random-profile-based 3D face recognition is presented. Although recent two-dimensional (2D) face recognition systems can achieve a reliable recognition rate under certain conditions, their performance is limited by internal and external changes, such as illumination and pose variation. To address these issues, 3D face recognition, which uses 3D face data, has recently received much attention. However, the performance of 3D face recognition highly depends on the precision of acquired 3D face data, while also requiring more computational power and storage capacity than 2D face recognition systems. In this paper, we present a developed nonintrusive 3D face modeling system composed of a stereo vision system and an invisible near-infrared line laser, which can be directly applied to profile-based 3D face recognition. We further propose a novel random-profile-based 3D face recognition method that is memory-efficient and pose-invariant. The experimental results demonstrate that the reconstructed 3D face data consists of more than 50 k 3D point clouds and a reliable recognition rate against pose variation.