Hang-Bong Kang

Affective content detection using HMMs

This paper discusses a new technique for detecting affective events using Hidden Markov Models(HMM). To map low level features of video data to high level emotional events, we perform empirical study on the relationship between emotional events and low-level features. After that, we compute simple low-level features that represent emotional characteristics and construct a token or observation vector by combining low level features. The observation vector sequence is tested to detect emotional events through HMMs. We create two HMM topologies and test both topologies. The affective events are detected from our proposed models with good accuracy.

Various Approaches for Driver and Driving Behavior Monitoring: A Review

In recent years, driver drowsiness and distraction have been important factors in a large number of accidents because they reduce driver perception level and decision making capability, which negatively affect the ability to control the vehicle. One way to reduce these kinds of accidents would be through monitoring driver and driving behavior and alerting the driver when they are drowsy or in a distracted state. In addition, if it were possible to predict unsafe driving behavior in advance, this would also contribute to safe driving. In this paper, we will discuss various monitoring methods for driver and driving behavior as well as for predicting unsafe driving behaviors. In respect to measurement methods of driver drowsiness, we discussed visual and non-visual features of driver behavior, as well as driving performance behaviors related to vehicle-based features. Visual feature measurements such as eye related measurements, yawning detection, facial expression are discussed in detail. As for non-visual features, we explore various physiological signals and possible drowsiness detection methods that use these signals. As for vehicle-based features, we describe steering wheel movement and the standard deviation of lateral position. To detect driver distraction, we describe head pose and gaze direction methods. To predict unsafe driving behavior, we explain predicting methods based on facial expressions and car dynamics. Finally, we discuss several issues to be tackled for active driver safety systems. They are 1) hybrid measures for drowsiness detection, 2) driving context awareness for safe driving, 3) the necessity for public data sets of simulated and real driving conditions.

Abnormal behavior detection using hybrid agents in crowded scenes

We categorize the behaviors of people into individual and group interactive behavior.We propose a hybrid agent system that includes static and dynamic agents in a scene.We represent the behavior of a crowd as a bag of words to detect abnormal behavior. In this paper, we propose a hybrid agent method to detect abnormal behaviors in a crowded scene. In real-life situations, abnormal behavior occurs by violent movement which is apparent as sudden speeding up, or chaotic movement in a restricted area, or movement contrasting with that of ones neighbors such as in a panic situation. In our model, we categorize the behaviors of people into individual behavior and group interactive behavior. Individual behavior is defined only by native motion information such as speed and direction. By contrast, group interactive behavior is defined by information concerning interactive motion between neighbors. We propose a hybrid agent system that includes static and dynamic agents to observe efficiently the corresponding individual and interactive behaviors in a crowded scene. The static agent is assigned to a specific spot and analyzes motion information near that spot. Unlike the static agent, the dynamic agent is assigned to a moving object and analyzes motion information of neighbors as well as oneself by following the objects movement. We represent the behavior of a crowd as a bag of words through the integration of static and dynamic agent information to determine abnormalities in the crowd behavior. The experimental results show that our proposed method efficiently detects abnormal behaviors in crowded scenes.

Constant Time Stereo Matching

Typically, local methods for stereo matching are fast but have relatively low degree of accuracy while global ones, though costly, achieve a higher degree of accuracy in retrieving disparity information. Recently, however, some local methods such as those based on segmentation or adaptive weights are suggested to possibly achieve more accuracy than global ones in retrieving disparity information. The problem for these newly suggested local methods is that they cannot be easily adopted since they may require more computational costs which increase in proportion to the window size they use. To reduce the computational costs, therefore, we propose in this paper the stereo matching method that use domain weight and range weight similar to those in the bilateral filter. Our proposed method shows constant time O(1) for the stereo matching. Our experiments spend constant time for computation regardless of the window size but our experimental results show that the accuracy of generated depth map is as good as the ones suggested by recent methods.

Multi-modal feature integration for secure authentication

In this paper, we propose a new multi-modal feature integration for secure authentication. We introduce behavioral information as well as biometrics information for the person of interest to test his verification. For continuous authentication, temporal score integration method is presented that incorporates biometrics and behavioral features. The proposed method was evaluated under several real situations and promising results were obtained.

Structural alterations in the prefrontal cortex mediate the relationship between Internet gaming disorder and depressed mood

Adaptive gaming use has positive effects, whereas depression has been reported to be prevalent in Internet gaming disorder (IGD). However, the neural correlates underlying the association between depression and Internet gaming remain unclear. Moreover, the neuroanatomical profile of the striatum in IGD is relatively less clear despite its important role in addiction. We found lower gray matter (GM) density in the left dorsolateral prefrontal cortex (DLPFC) in the IGD group than in the Internet gaming control (IGC) group and non-gaming control (NGC) group, and the GM density was associated with lifetime usage of Internet gaming, depressed mood, craving, and impulsivity in the gaming users. Striatal volumetric analysis detected a significant reduction in the right nucleus accumbens (NAcc) in the IGD group and its association with lifetime usage of gaming and depression. These findings suggest that alterations in the brain structures involved in the reward system are associated with IGD-related behavioral characteristics. Furthermore, the DLPFC, involved in cognitive control, was observed to serve as a mediator in the association between prolonged gaming and depressed mood. This finding may provide insight into an intervention strategy for treating IGD with comorbid depression.

Changes in ocular factors according to depth variation and viewer age after watching a three-dimensional display.

Objective To investigate changes in ocular factors according to the binocular disparity in three-dimensional (3D) images and age after watching 3D display. Methods A total of 38 volunteers were enrolled, and they watched a 3D display with a 1° or 3° disparity for 30 min at an interval of 1 week. The near point of accommodation (NPA), near point of convergence (NPC) and tear break-up time (tBUT) of each subject were measured before and after watching the 3D display. In addition, the tear meniscus height and depth were measured using Visante optical coherence tomography and tear osmolarity was measured using TearLab osmometer. A survey of subjective symptoms was also conducted. Results NPA and NPC increased after watching the 3D display (p<0.05). NPC and NPA increased more in the 40s–50s group (ie, subjects aged in their 40s and 50s) than in the 20s–30s group (ie, subjects aged in their 20s and 30s) after watching 3D content with a 3° disparity (p<0.05). tBUT and tear meniscus height and depth decreased after watching 3D content (p<0.05). They decreased more in the 40s–50s group than in the 20s–30s group after watching 3D content with a 3° disparity (p<0.05). Recovery times of NPA and NPC were significantly greater after watching 3D content with a 3° disparity and in the 40s–50s group (p<0.05). Conclusions Watching a 3D display affects accommodation and convergence abilities and tear dynamics in a transient fashion, especially in the case of 3D images with a large binocular disparity, and in older subjects. These results provide helpful information for establishment of guidelines for 3D equipment manufacturers.

Prediction of crime occurrence from multi-modal data using deep learning

In recent years, various studies have been conducted on the prediction of crime occurrences. This predictive capability is intended to assist in crime prevention by facilitating effective implementation of police patrols. Previous studies have used data from multiple domains such as demographics, economics, and education. Their prediction models treat data from different domains equally. These methods have problems in crime occurrence prediction, such as difficulty in discovering highly nonlinear relationships, redundancies, and dependencies between multiple datasets. In order to enhance crime prediction models, we consider environmental context information, such as broken windows theory and crime prevention through environmental design. In this paper, we propose a feature-level data fusion method with environmental context based on a deep neural network (DNN). Our dataset consists of data collected from various online databases of crime statistics, demographic and meteorological data, and images in Chicago, Illinois. Prior to generating training data, we select crime-related data by conducting statistical analyses. Finally, we train our DNN, which consists of the following four kinds of layers: spatial, temporal, environmental context, and joint feature representation layers. Coupled with crucial data extracted from various domains, our fusion DNN is a product of an efficient decision-making process that statistically analyzes data redundancy. Experimental performance results show that our DNN model is more accurate in predicting crime occurrence than other prediction models.

Integrated multiple behavior models for abnormal crowd behavior detection

We propose a method to detect abnormal crowd behavior using integrated multiple behavior models. Traditional abnormal detection methods are based only on personal behavior models. However, a single behavior model cannot accurately reflect complex crowd behavior. To solve this problem, we introduce an integrated multiple behavior model for accurate abnormal behavior detection in a complex crowd scene. We use not only the personal behavior model, but also multiple social behavior models. The experimental results show that our proposed method efficiently detects the abnormal behavior in a crowded scene.

Text sentiment classification for SNS-based marketing using domain sentiment dictionary

In this paper, we propose a new method of classifying the sentiment behind tweets that contains formal and informal vocabulary. Previous methods used only formal vocabulary to classify the sentiments behind the sentences. However, these methods are ineffective in classifying texts since internet users make sentences using informal vocabulary. In addition, we use emotion based vocabulary to classify the sentiment behind texts. Feature vectors extracted from the vocabulary are classified by Support Vector Machine (SVM). Our proposed method shows a strong performance in the classifying the emotion behind the text.