Jun-Su Kang

Modulation of resource allocation by intelligent individuals in linguistic, mathematical and visuo-spatial tasks

This study investigates two questions: first, how individuals with high-intelligence allocate cognitive resources while solving linguistic, mathematical and visuo-spatial tasks with varying degree of difficulty as compared to individuals with low intelligence? Second, how to distinguish between high and low intelligent individuals by analyzing pupil dilation and eye blink together? We measured the response time, error rates along with pupil dilation and eye blink rate that indicate resource allocation. We divided the whole processing into three stages namely: pre-stimuli (5s prior to stimuli onset), during stimuli and post stimuli (until 5s after the response) for better assessment of preparation and resource allocation strategies. Individuals with high intelligence showed greater task evoked pupil dilation, decreased eye blink with less response time and error rates during-stimuli stage (processing) of tough linguistic and visuo-spatial tasks but not during mathematical tasks. The finding suggests that individuals with high intelligence allocate more resources if the task demands are high else they allocate less resources. Greater pre-stimuli pupil dilation and increased eye blink of high intelligent individuals in all tasks indicated their attentiveness and preparedness. The result of our study shows that individuals with high intelligence are more attentive and flexible in terms of altering the resource allocation strategy according to task demand. Eye-blinks along with pupil dilation and other behavioral parameters can be reliably used to assess the intelligence of an individual and the analysis of pupil dilation and blink rate at pre-stimuli stage can be crucial in distinguishing individuals with varying intelligence.

Emotion classification in movie clips based on 3D fuzzy GIST and EEG signal analysis

In this paper, we propose an emotion classification model, which can differentiate human-like emotions by using visual and electroencephalography (EEG) dynamic features. To understand human emotions in a more natural situation, we use dynamic stimuli such as movies for the analysis. We incorporate the 3D fuzzy GIST to effectively describe both dynamic visual features and EEG signals. The extracted features are used as inputs to an adaptive neuro-fuzzy inference system (ANFIS). The classifier is provided with the mean opinion scores as the teaching signals. Experimental results show that the system using both low-level visual feature and semantic level EEG feature not only discriminates the positive emotional features from the negative ones but also can get the more stable result than the model using only visual or EEG information.

Advanced driver assistant system based on monocular camera

An estimated 20-50 million people get injured and at least 1.2 million people die in automotive accidents every year. To avoid this, several security systems such as lane departure warning system, vehicle detection system, advanced cruise control system, etc., have been developed. But they are very expensive. In our research, we focus on developing low cost intelligent vehicle systems. In this paper, we propose a driver assistant system based on monocular camera. The proposed system consists of two parts. First part is vehicle detection module based on adaptive boosting using extent modified census transform (EMCT). Second part is lane detection module using RANdom Sample Consensus (RANSAC). Experimental results of system shows that both modules perform robustly with less computational load and that it can be a reliable intelligent vehicle system.

Analysis of Cognitive Load for Language Processing Based on Brain Activities

The present study attempted to investigate the role of memory or cognitive load in language processing using an EEG. Twelve healthy righthanded male adults were asked to read a story twice and their brain activities were recorded using an EEG: (i) focusing on meaning of the content only (M) and (ii) focusing on both meaning and form or grammar (M+F). The results demonstrated significant differences in upper alpha and upper beta bands according to reading instructions, which indicates different degrees of cognitive load. The findings make a significant contribution to language acquisition in that they offer valuable information regarding memory and cognitive load in language processing. Thus, they help language researchers and educators in the field of second language acquisition (SLA) develop more effective ways of instructional design and in turn lead their students to better learning outcomes.

Stress status classification based on EEG signals

In daily life, humans get stress very often. Stress is one of the important factors of healthy life and closely related to the quality of life. Too much stress is known to cause hormone imbalance of our body, and it is observed by the brain and bio signals. Based on this, the relationship between brain signal and stress is explored, and brain signal based stress index is proposed in our work. In this study, an EEG measurement device with 32 channels is adopted. However, only two channels (FP1, FP2) are used to this study considering the applicability of the proposed method in real enveironment, and to compare it with the commercial 2 channel EEG device. Frequency domain features are power of each frequency bands, subtraction, addition, or division by each frequency bands. Features in time domain are hurst exponent, correlation dimension, lyapunov exponent, etc. Total 6 subjects are participated in this experiment with English sentence reading task given. Among several candidate features, shows the best test performance (70.8%). For future work, we will confirm the results is consistent in low price EEG device.

Convolutional Neural Networks Considering Robustness Improvement and Its Application to Face Recognition

This paper proposes a novel activation function to promote robustness to the outliers of the training samples. Data samples in the decision boundaries are weighted more by adding the derivatives of the sigmoid function outputs to avoid drastic update of the network weights. Therefore, the network becomes more robust to outliers and noisy patterns. We also present appropriate backpropagation learning algorithm for the convolutional neural networks. We evaluate the performance improvement by the proposed method on a face recognition task, and proved that it outperformed the state of art face recognition methods.

Concentration Monitoring with High Accuracy but Low Cost EEG Device

Concentration is an important part of our life especially during learning or thinking. Visually or auditory evoked concentration affects information processing in human brain. To understand the concentration process of humans, the underlying neural mechanism needs to be explored. EEG device is a promising device to understand underlying neural mechanism of various cognitive functions. In this paper, we propose an accurate concentration monitoring method using a low cost EEG device. Our low cost EEG device has two channel electrodes (FP1, FP2). Usually small channel EEG devices face filtering problem because commonly used filtering method, such as ICA, fails with less number of electrodes. In our work, we investigate effective filters for removing noises from raw data and suitable features for monitoring the concentration status with the low cost EEG device in real time. We collect EEG data from 10 participants for rest state with open eyes and concentration task state. For concentration task, Sudoku game is used. Using support vector machine, we successfully distinguish between rest state and concentration state over 88 % accuracy in real time.

EEG Based Coherence Analysis for Identifying Inter Individual Differences in Language and Logic Study

According to Multiple Intelligences MI theory by Howard Gardner, humans intelligence can be divided into linguistic, logical/mathematical, musical, spatial, bodily/kinesthetic, interpersonal, intrapersonal and naturalistic areas. In this paper, two groups of students with high and low MI scores in each area for linguistic and logical/mathematical intelligences were categorized and tested based on a questionnaire in an experiment. Electroencephalogram EEG was recorded from 16 electrodes using Emotiv EPOC for 29 kindergarten children and 9 high school students during the assessment. To investigate the neurophysiological substrates of intelligence, EEG signal based coherence analysis in alpha, beta, gamma and theta frequency bands was performed. Our findings indicate that the group with low MI score for both language and logic showed wider and distributed cognitive activations suggesting an increased effort in processing a particular task. The group with high MI score for language showed effective connectivity within left-hemisphere and low activation in the right parietal lobe. The group with high MI score for logic/mathematics showed increased frontal activation. Performance in language and logic test was further correlated with effective connectivity in the task specific areas of brain. Based on our results we conclude that a smarter brain for language and logic is associated with the limited but affective connectivity.