• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.15 no.6
• /
• pp.2098-2114
• /
• 2021

Light intensity variation is one of the key factors which affect the accuracy of vehicle face re-identification, so in order to improve the robustness of vehicle face features to light intensity variation, a Nonnegative Matrix Factorization model with the constraint of image acquisition time difference is proposed. First, the original features vectors of all pairs of positive samples which are used for training are placed in two original feature matrices respectively, where the same columns of the two matrices represent the same vehicle; Then, the new features obtained after decomposition are divided into stable and variable features proportionally, where the constraints of intra-class similarity and inter-class difference are imposed on the stable feature, and the constraint of image acquisition time difference is imposed on the variable feature; At last, vehicle face matching is achieved through calculating the cosine distance of stable features. Experimental results show that the average False Reject Rate and the average False Accept Rate of the proposed algorithm can be reduced to 0.14 and 0.11 respectively on five different datasets, and even sometimes under the large difference of light intensities, the vehicle face image can be still recognized accurately, which verifies that the extracted features have good robustness to light variation.

• KIPS Transactions on Software and Data Engineering
• /
• v.4 no.7
• /
• pp.291-296
• /
• 2015

In this paper, we propose a new confidence measure using pixel selection for eye detection and design a hybrid eye detector. For this, we produce sub-images by applying a pixel selection method to the eye patches and construct the BDA(Biased Discriminant Analysis) feature space for measuring the confidence of the eye detection results. For a hybrid eye detector, we select HFED(Haar-like Feature based Eye Detector) and MFED(MCT Feature based Eye Detector), which are complementary to each other, as basic detectors. For a given image, each basic detector conducts eye detection and the confidence of each result is estimated in the BDA feature space by calculating the distances between the produced eye patches and the mean of positive samples in the training set. Then, the result with higher confidence is adopted as the final eye detection result and is used to the face alignment process for face recognition. The experimental results for various face databases show that the proposed method performs more accurate eye detection and consequently results in better face recognition performance compared with other methods.

• Convergence Security Journal
• /
• v.18 no.5_1
• /
• pp.59-66
• /
• 2018

The performance of a network intrusion detection system using the machine learning method depends heavily on the composition and the size of the feature set. The detection accuracy, such as the detection rate or the false positive rate, of the system relies on the feature composition. And the time it takes to train and detect depends on the size of the feature set. Therefore, in order to enable the system to detect intrusions in real-time, the feature set to beused should have a small size as well as an appropriate composition. In this paper, we show that the size of the feature set can be further reduced without decreasing the detection rate through using Pearson correlation coefficient between features along with the multi-objective genetic algorithm which was used to shorten the size of the feature set in previous work. For the evaluation of the proposed method, the experiments to classify 10 kinds of attacks and benign traffic are performed against NSL_KDD data set.

• Journal of Institute of Control, Robotics and Systems
• /
• v.15 no.5
• /
• pp.457-464
• /
• 2009

The weak classifier of AdaBoost algorithm is a central classification element that uses a single criterion separating positive and negative learning candidates. Finding the best criterion to separate two feature distributions influences learning capacity of the algorithm. A common way to classify the distributions is to use the mean value of the features. However, positive and negative distributions of Haar-like feature as an image descriptor are hard to classify by a single threshold. The poor classification ability of the single threshold also increases the number of boosting operations, and finally results in a poor classifier. This paper proposes a weak classifier that uses multiple criterions by adding a probabilistic criterion of the positive candidate distribution with the conventional mean classifier: the positive distribution has low variation and the values are closer to the mean while the negative distribution has large variation and values are widely spread. The difference in the variance for the positive and negative distributions is used as an additional criterion. In the learning procedure, we use a new classifier that provides a better classifier between them by selective switching between the mean and standard deviation. We call this new type of combined classifier the "Mixed Weak Classifier". The proposed weak classifier is more robust than the mean classifier alone and decreases the number of boosting operations to be converged.

• Smart Media Journal
• /
• v.9 no.3
• /
• pp.9-17
• /
• 2020

Hippocampus is a complex brain structure embedded deep into the temporal lobe. Studies have shown that this structure gets affected by neurological and psychiatric disorders and it is a significant landmark for diagnosing neurodegenerative diseases. Hippocampus features play very significant roles in region-of-interest based analysis for disease diagnosis and prognosis. In this study, we have attempted to learn the embeddings of this important biomarker. As conventional metric learning methods for feature embedding is known to lacking in capturing semantic similarity among the data under study, we have trained deep Siamese convolutional neural network for learning metric of the hippocampus. We have exploited Gwangju Alzheimer's and Related Dementia cohort data set in our study. The input to the network was pairs of three-view patches (TVPs) of size 32 × 32 × 3. The positive samples were taken from the vicinity of a specified landmark for the hippocampus and negative samples were taken from random locations of the brain excluding hippocampi regions. We have achieved 98.72% accuracy in verifying hippocampus TVPs.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.5 no.10
• /
• pp.1751-1768
• /
• 2011

Falls are one of the most concerned accidents for elderly people and often result in serious physical and psychological consequences. Many researchers have studied fall detection techniques in various domain, however none released to a commercial product satisfying user requirements. We present a systematic modeling and evaluating procedure for best classification performance and then do experiments for comparing the performance of six procedures to get a statistical classifier based wrist-type fall detector to prevent dangerous consequences from falls. Even though the wrist may be the most difficult measurement location on the body to discern a fall event, the proposed feature deduction process and fall classification procedures shows positive results by using data sets of fall and general activity as two classes.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.9 no.8
• /
• pp.3169-3181
• /
• 2015

Nowadays, online word-of-mouth has become a powerful influencer to marketing and sales in business. Opinion mining and sentiment analysis is frequently adopted at market research and business analytics field for analyzing word-of-mouth content. However, there still remain several challengeable areas for 1) sentiment analysis aiming for Korean word-of-mouth content in film market, 2) availability of machine learning models only using linguistic features, 3) effect of the size of the feature set. This study took a sample of 10,000 movie reviews which had posted extremely negative/positive rating in a movie portal site, and conducted sentiment analysis with four machine learning algorithms: naïve Bayesian, decision tree, neural network, and support vector machines. We found neural network and support vector machine produced better accuracy than naïve Bayesian and decision tree on every size of the feature set. Besides, the performance of them was boosting with increasing of the feature set size.

• Journal of Institute of Control, Robotics and Systems
• /
• v.16 no.5
• /
• pp.469-475
• /
• 2010

This paper presents a method of classifying discolored defects of metal pads using geometric texture for AFVI (Automated Final Vision Inspection) systems. In PCB manufacturing process, the metal pads on PCB can be oxidized and discolored partly due to various environmental factors. Nowadays the discolored defects are manually detected and rejected from the process. This paper proposes an efficient geometric texture feature, SUTF (Symmetry and Uniformity Texture Feature) based on the symmetric and uniform textural characteristics of the surface of circular metal pads for automating AFVI systems. In practical experiments with real samples acquired from a production line, 30 discolored images and 1232 roughness images are tested. The experimental results demonstrate that the proposed method using SUTFs provides better performance compared to Gabor feature with 0% FNR (False Negative Rate) and 1.46% FPR (False Positive Rate). The performance of the proposed method shows its applicability in the real manufacturing systems.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2015.04a
• /
• pp.779-781
• /
• 2015

최근 여러 매체에서 화두가 되고 있는 자율 주행 자동차나 Advanced driver assistance systems (ADAS)과 같은 분야에서 보행자 검출 기술은 핵심 요소 기술 중에 하나로 손꼽히고 있다. 특히, 인간의 인지 부하(Cognitive Load)를 고려했을 때, 주행 중에 발생할 수 있는 모든 사건을 다룬다는 것은 매우 어렵기 때문에, 앞서 언급한 방법의 도움을 받아 도로 주행 중에 발생 될 수 있는 인명 사고율을 줄이고자 하는데 그 목적이 있다. 본 논문에서는 Integral Channel Feature를 사용하여 AdaBoost 알고리즘으로 보행자 검출을 위한 분류기를 구현하였다. 그 결과, INRIA에서 제공되는 Pedestrian dataset에서 Detection rate는 97%이상, False positive는 1%에 정도로 나타났다.

• IEMEK Journal of Embedded Systems and Applications
• /
• v.6 no.5
• /
• pp.295-302
• /
• 2011

This paper proposes facial feature tracking based on modified ART2 neural networks. And we also suggest new measurement devices such as 'Persistence Exponent' and 'Moving Space Exponent' for the criterion of input vector which consists features. The proposed methods have been applied to classify 48 students by 2-class (ADHD positive, ADHD negative). The results of the experiment have shown that the proposed methods are effective for ADHD Behavior Pattern Classification based on the Image Processing.