Hakjae Kim

Depth-Based Detection of Standing-Pigs in Moving Noise Environments

In a surveillance camera environment, the detection of standing-pigs in real-time is an important issue towards the final goal of 24-h tracking of individual pigs. In this study, we focus on depth-based detection of standing-pigs with “moving noises”, which appear every night in a commercial pig farm, but have not been reported yet. We first apply a spatiotemporal interpolation technique to remove the moving noises occurring in the depth images. Then, we detect the standing-pigs by utilizing the undefined depth values around them. Our experimental results show that this method is effective for detecting standing-pigs at night, in terms of both cost-effectiveness (using a low-cost Kinect depth sensor) and accuracy (i.e., 94.47%), even with severe moving noises occluding up to half of an input depth image. Furthermore, without any time-consuming technique, the proposed method can be executed in real-time.

A fingerprint-based user authentication protocol considering both the mobility and security in the telematics environment

With the advance of the Internet and mobile communication techniques, the telematics environment where users in vehicles can use the Internet service has been realized. For the safe driving, however, we propose that user authentication for the Internet service is performed by using the drivers fingerprint, instead of typing his/her password. Since the drivers fingerprint is private information to be protected and the size of the fingerprint information is much larger than that of a typical password, we need a different user authentication protocol for the telematics environment. That is, in addition to the compliance with the standard X9.84 protocol to protect the fingerprint information transmitted, we use the watermarking technique to lessen the privacy threat, and propose a secure and efficient protocol between Access Points (APs) considering the possible hand-off during the authentication in the mobile telematics environment. Based on the experimental measurement of the proposed protocol, we confirm that the fingerprint-based user authentication can be performed in real-time in the telematics environment.

Heterogeneous computing for a real-time pig monitoring system

Video sensor data has been widely used in automatic surveillance applications. In this study, we present a method that automatically detects pigs in a pig room by using depth information obtained from a Kinect sensor. For a real-time implementation, we propose a means of reducing the execution time by applying parallel processing techniques. In general, most parallel processing techniques have been used to parallelize a specific task. In this study, we consider parallelization of an entire system that consists of several tasks. By applying a scheduling strategy to identify a computing device for each task and implementing it with OpenCL, we can reduce the total execution time efficiently. Experimental results reveal that the proposed method can automatically detect pigs using a CPU-GPU hybrid system in real time, regardless of the relative performance between the CPU and GPU.

Lying-Pig Detection using Depth Information

In a surveillance camera environment, detecting pigs in a pig room is an important issue in terms of automatic management. In addition, it is important to analyze the area of lying-pigs in order to estimate the temperature of a pig room. However, it is difficult to recognize whether the pigs are lying or standing from RGB data because the data do not have any height information. In this study, we propose a method to detect lying-pigs by using depth information obtained from a 3D Kinect camera. According to the experimental results, we confirmed that the proposed method could detect lying-pigs by using a difference of the depth information between pigs and floor.

Touching-Pigs Segmentation using Concave Points in Continuous Video Frames

Caring individual pigs in large-scale pig farms is an important issue for preventing infectious diseases. Accordingly, many researchers have been researched about group-housed pig monitoring systems. However, it is challenging to identify individual pigs because the systems misidentify touching-pigs as a single pig. In this paper, we solve the touching-pig problem by using concave points of continuous video frames. We interpret a two dimensional outline data as a one dimensional time-series data of touching pigs and align the time-series data of continuous video frames. The experimental results show that the proposed method can segment the touching-pigs more accurate than generally used methods in real time.

Energy-Efficient Transmissions of Fingerprint Images Using Both Encryption and Fingerprinting Techniques

As user authentication by using biometric information such as fingerprint has been widely accepted, there has been a growing interest in protecting the biometric information itself against external attackers. Especially, as the mobile handheld devices equipped with fingerprint sensors are produced, it becomes important to protect the private information of an user(i.e., fingerprint image) in the remote applications. In this paper, we propose an energy-efficient technique to transmit fingerprint images from the handheld device to a server securely using both encryption and fingerprinting. In addition to the encryption to guarantee the confidentiality of the fingerprint image transmitted, we insert the logo of the application into the fingerprint image(i.e., fingerprinting) to indicate to whom the fingerprint image is transmitted(i.e., a possible privacy violator). In particular, to lower the computational burden of the handheld device, we not only offload some heavy computation to the server without compromising the security, but also apply some techniques to the handheld device such as loop transformation and data compaction. Based on experimental results, we confirm that our technique can transmit fingerprint images from the handheld device in an energy-efficient way(by a factor of 2) and detect the intentional/unintentional leakage of the transmitted fingerprint image from the server – privacy violator.