Hae-Min Moon

A Survey and Proposed Framework on the Soft Biometrics Technique for Human Identification in Intelligent Video Surveillance System

Biometrics verification can be efficiently used for intrusion detection and intruder identification in video surveillance systems. Biometrics techniques can be largely divided into traditional and the so-called soft biometrics. Whereas traditional biometrics deals with physical characteristics such as face features, eye iris, and fingerprints, soft biometrics is concerned with such information as gender, national origin, and height. Traditional biometrics is versatile and highly accurate. But it is very difficult to get traditional biometric data from a distance and without personal cooperation. Soft biometrics, although featuring less accuracy, can be used much more freely though. Recently, many researchers have been made on human identification using soft biometrics data collected from a distance. In this paper, we use both traditional and soft biometrics for human identification and propose a framework for solving such problems as lighting, occlusion, and shadowing.

Automatic lung segmentation for large-scale medical image management

Digital medical images assist specialists in improving their diagnostic efficiency and in treating diseases. For example, the chest Computed Tomography (CT) images help in diagnosing the lung disease. The chest CT scan generates multiple images of a patient’s lung. The size of the medical imaging data has increased with the usage of medical images. In a picture archiving and communication system, large-scale medical images must be transmitted to specialists through either wired or wireless communications and retained in the archive. Hence, medical images have to be compressed, and there should be no damage to the Region of Interest (RoI) during compression. In order to protect the RoI, image segmentation is needed to detect RoI in medical images. Among the various image segmentation methods available, the method using Level-set is robust to irregular noises. However, the problems faced in using this method include manual input of the initial contour and slow performance speed. Inputting an initial contour to the Level-set that correctly fits the object’s form helps in reducing the number of repetitions. This in turn helps in improving the segmentation performance speed. However, it is difficult for a user to input an appropriate initial contour. Therefore, this paper aims at providing a method to auto-configure the initial contour in the Level-set method. Multi-resolution analysis helps in reducing the pace of the auto-configuration process of the initial contour. In addition, the volume data of a CT image is used to prevent data loss that occurs during the MRA transformation process. Studies have confirmed that the proposed method facilitates drastic improve.

Privacy Protection Technology in Video Surveillance System

Since the increase of terrors and crimes, the demand of the video surveillance system is increasing. The video surveillance system has been operated for public interest by the police and local government. However private information such as faces or behavior patterns can be recorded in CCTV, it may cause the invasion of privacy. This paper proposes a video surveillance system that satisfies both functions of privacy protection and surveillance. The proposed system operates multiple surveillance cameras with a single RFID system. The RFID system is used in order to recognize internal or external personnel of object at entrance and cameras are used to extract features such as height and clothing-colors of the object. The proposed system can reduce invasion of privacy by intensity of scrambling according to access levels. Based on the experimental results, we are able to confirm that surveillant can classify personnel information fast and exactly in the display.

JPEG 2000 and Digital Watermarking Technique Using in Medical Image

The Picture Archiving and Communication System(PACS) was introduced for computerization of the medical system and telediagnosis between the hospital. It is becoming possible to create, store, and transmit medical images via PACS. There has been a growing interest in protecting medical images with an enormous amount of information. To improve transmission speed among the hospitals, the medical image should be compression JPEG 2000 by high compression ratio. This paper proposes an algorithm that utilizes both JPEG 2000 and robust watermarking for protection and compression of the medical image. With the proposed algorithm, it takes considerably less time to do JPEG 2000 and watermarking than when they are done separately. Based on the experiment results, it takes 0.72 second for the proposed algorithm and 1.11 second when they are done separately. We confirmed that the proposed algorithm was faster than when they are done separately.

Implementation of the Privacy Protection in Video Surveillance System

Due to increased terrors and crimes, the use of the video surveillance camera system is increasing. It has been operated for public interest such as prevention of crimes and fly-tipping by the police and local government, but private information such as faces or behavior patterns can be recorded in CCTV. When the recorded video data is exposed, it may cause an invasion to privacy and crimes. This paper analyses conventional methods of privacy protection in surveillance camera systems and applied scrambling and RFID system to existing surveillance systems to prevent privacy exposure in monitoring simultaneously for both privacy protection and surveillance. The proposed system adjusts the intensities of privacy according to access levels to reduce invasion of privacy by people who are not concerned.

Intelligent video surveillance system using two-factor human information

Recently, with the increase of terror and crime the utilization of security surveillance systems including CCTV increases. However, the issues of privacy invasion occurred by exposing the data recorded through video surveillance system have been raised. In this paper, the intelligent video surveillance system which can prevent the invasion of privacy and complement the monitoring function which will be declined by privacy protection is proposed. The proposed system checks the identification using the smart card and camera at the entrance and extracts the height and color information of object. Inside of the building where only a camera is installed without smart card terminal installed checks the identification using height and color information of object. With the implemented identification function the proposed system sorts the object receiving the privacy protection, implements the monitoring function and provides monitoring and protection function at the same time.

A study on the digital watermarking technique for JPEG 2000

Recently, according to the number of Internet in widely use and the development of the related application, the distribution and use of multimedia contents, such as digital images, are very easy. Since these can modify as easy as use, multimedia contents need newly compressed JPEG 2000 and digital watermarking technique because of its enormous amount of information. To protect the digital image in real-time, this paper proposes two methods that Dual Watermarking Method (DWM) based JPEG 2000 and performing JPEG 2000 and watermarking synchronously. DWM consists of the robust and fragile watermarking. Also, in DWM, watermarks are carefully embedded avoiding the areas of Region Of Interest (ROI) and the edge of the digital image to protect the integrity of it. In performing JPEG 2000 and watermarking, it takes considerably less time to do JPEG 2000 and watermarking than when they are done separately. Based on the experimental results, it takes 0.72 second for JPEG 2000 and watermarking, and 1.11 second when they are done separately.

Auto-configuration of the Initial Contour Using Level-Set Method for Lung Segmentation on Chest CT Images

Digital medical images assist specialists in improving their diagnostic efficiency and in treating diseases. For example, the chest Computed Tomography (CT) images help in diagnosing the lung disease. The chest CT scan generates multiple images of a patient’s lung. A medical image processing technique helps in segmenting these images. It is important to perform this step before processing other medical images. Among the various image segmentation methods available, the method using Level-set is robust to irregular noises. However, the problems faced in using this method include manual input of the initial contour and slow performance speed. Inputting an initial contour to the Level-set that correctly fits the object’s form helps in reducing the number of repetitions. This in turn helps in improving the segmentation performance speed. However, it is difficult for a user to input an appropriate initial contour. Therefore, this paper aims at providing a method to auto-configure the initial contour in the Level-set method. Multi-Resolution Analysis (MRA) helps in reducing the pace of the auto-configuration process of the initial contour. In addition, the volume data of a CT image is used to prevent data loss that occurs during the MRA transformation process. Studies have confirmed that the proposed method facilitates drastic improvements in the performance time and in the segmentation results of chest CT images.

A novel double linear-cubic convolution interpolation for digital image scaling

As the better high quality image is required for digital image scaling, longer processing time is required so the technology that can make the high quality image quickly is needed. We propose the double linear--cubic convolution interpolation creating the high quality image with low complexity. When compared to peak signal-to-noise ratio(PSNR) and computation time, the proposed interpolation provided better PSNR and low complexity than bicubic convolution interpolation.

The PCA-Based long distance face recognition using multiple distance training images for intelligent surveillance system

In this paper, PCA-based long distance face recognition algorithm applicable to the environment of intelligent video surveillance system is proposed. While the existing face recognition algorithm uses the short distance images for training images, the proposed algorithm uses face images by distance extracted from 1m to 5m for training images. Face images by distance, which are used for training images and test images, are normalized through bilinear interpolation. The proposed algorithm has improved face recognition performance by 4.8% in short distance and 16.5% in long distance so it is applicable to the intelligent video surveillance system.