• Journal of KIISE:Software and Applications
• /
• v.36 no.7
• /
• pp.548-559
• /
• 2009

The cascade face detector learned by Adaboost algorithm, which was proposed by Viola and Jones, is state of the art face detector due to its great speed and accuracy. In spite of its great performance, it still suffers from false alarms, and more computation is required to reduce them. In this paper, we want to reduce false alarms with less computation using facial color. Using facial color information, proposed face detection model scans sub-window efficiently and adapts a fast face/non-face classifier at the first stage of cascade face detector. This makes face detection faster and reduces false alarms. For facial color filtering, we define a facial color membership function, and facial color filtering image is obtained using that. An integral image is calculated from facial color filtering image. Using this integral image, its density of subwindow could be obtained very fast. The proposed scanning method skips over sub-windows that do not contain possible faces based on this density. And the face/non-face classifier at the first stage of cascade detector rejects a non-face quickly. By experiment, we show that the proposed face detection model reduces false alarms and is faster than the original cascade face detector.

• Journal of KIISE:Software and Applications
• /
• v.29 no.4
• /
• pp.270-277
• /
• 2002

Text filtering is a task of deciding whether a document has relevance to a specified topic. As Internet and Web becomes wide-spread and the number of documents delivered by e-mail explosively grows the importance of text filtering increases as well. The aim of this paper is to improve the accuracy of text filtering systems by using machine learning techniques. We apply AdaBoost algorithms to the filtering task. An AdaBoost algorithm generates and combines a series of simple hypotheses. Each of the hypotheses decides the relevance of a document to a topic on the basis of whether or not the document includes a certain word. We begin with an existing AdaBoost algorithm which uses weak hypotheses with their output of 1 or -1. Then we extend the algorithm to use weak hypotheses with real-valued outputs which was proposed recently to improve error reduction rates and final filtering performance. Next, we attempt to achieve further improvement in the AdaBoost's performance by first setting weights randomly according to the continuous Poisson distribution, executing AdaBoost, repeating these steps several times, and then combining all the hypotheses learned. This has the effect of mitigating the ovefitting problem which may occur when learning from a small number of data. Experiments have been performed on the real document collections used in TREC-8, a well-established text retrieval contest. This dataset includes Financial Times articles from 1992 to 1994. The experimental results show that AdaBoost with real-valued hypotheses outperforms AdaBoost with binary-valued hypotheses, and that AdaBoost iterated with random weights further improves filtering accuracy. Comparison results of all the participants of the TREC-8 filtering task are also provided.

• Transactions of the Korean Society of Automotive Engineers
• /
• v.17 no.1
• /
• pp.1-11
• /
• 2009

This paper proposes an algorithm capable of detecting vehicles in front or in rear using a monocular camera installed in a vehicle. The vehicle detection has been regarded as an important part of intelligent vehicle technologies. The proposed algorithm is mainly composed of two parts: 1)hypothesis generation of vehicles, and 2)hypothesis verification. The hypotheses of vehicles are generated by the analysis of vertical and horizontal edges and the detection of symmetry axis. The hypothesis verification, which determines vehicles among hypotheses, is done by the AdaBoost algorithm. The proposed algorithm is proven to be effective through experiments performed on various images captured on the roads.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.18 no.7
• /
• pp.655-661
• /
• 2017

It is a very important task to extract character regions contained in various input color images, because characters can provide significant information representing the content of an image. In this paper, we propose a new method for extracting character regions from various input images using MCT features and an AdaBoost algorithm. Using geometric features, the method extracts actual character regions by filtering out non-character regions from among candidate regions. Experimental results show that the suggested algorithm accurately extracts character regions from input images. We expect the suggested algorithm will be useful in multimedia and image processing-related applications, such as store signboard detection and car license plate recognition.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.17 no.11
• /
• pp.403-409
• /
• 2016

Recently, harmful content (such as images and photos of nudes) has been widely distributed. Therefore, there have been various studies to detect and filter out such harmful image content. In this paper, we propose a new method using Haar-like features and an AdaBoost algorithm for robustly extracting navel areas in a color image. The suggested algorithm first detects the human nipples through color information, and obtains candidate navel areas with positional information from the extracted nipple areas. The method then selects real navel regions based on filtering using Haar-like features and an AdaBoost algorithm. Experimental results show that the suggested algorithm detects navel areas in color images 1.6 percent more robustly than an existing method. We expect that the suggested navel detection algorithm will be usefully utilized in many application areas related to 2D or 3D harmful content detection and filtering.

• Transactions of the Korean Society of Automotive Engineers
• /
• v.24 no.2
• /
• pp.214-224
• /
• 2016

This paper proposes an algorithm to effectively detect the traffic lights and recognize the traffic signals using a monocular camera mounted on the front windshield glass of a vehicle in day time. The algorithm consists of three main parts. The first part is to generate the candidates of a traffic light. After conversion of RGB color model into HSI and YCbCr color spaces, the regions considered as a traffic light are detected. For these regions, edge processing is applied to extract the borders of the traffic light. The second part is to divide the candidates into traffic lights and non-traffic lights using Haar-like features and Adaboost algorithm. The third part is to recognize the signals of the traffic light using a template matching. Experimental results show that the proposed algorithm successfully detects the traffic lights and recognizes the traffic signals in real time in a variety of environments.

• KIISE Transactions on Computing Practices
• /
• v.23 no.7
• /
• pp.408-416
• /
• 2017

This paper proposes a real-time side-rear vehicle detection algorithm that detects vehicles quickly and accurately in blind spot areas when driving. The proposed algorithm uses a cascade classifier created by AdaBoost Learning using the MCT (modified census transformation) feature vector. Using this classifier, the smaller the detection window, the faster the processing speed of the MCT classifier, and the larger the detection window, the greater the accuracy of the MCT classifier. By considering these characteristics, the proposed algorithm uses two classifiers with different detection window sizes. The first classifier quickly generates candidates with a small detection window. The second classifier accurately verifies the generated candidates with a large detection window. Furthermore, the vehicle classifier and the wheel classifier are simultaneously used to effectively detect a vehicle entering the blind spot area, along with an adjacent vehicle in the blind spot area.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.20 no.6
• /
• pp.806-813
• /
• 2010

Human facial expression shows human's emotion most exactly, so it can be used as the most efficient tool for delivering human's intention to computer. For fast and exact recognition of human's facial expression on a 2D image, this paper proposes a new method which integrates an Discrete Adaboost classification algorithm and a neural network based recognition algorithm. In the first step, Adaboost algorithm finds the position and size of a face in the input image. Second, input detected face image into 5 Adaboost strong classifiers which have been trained for each facial expressions. Finally, neural network based recognition algorithm which has been trained with the outputs of Adaboost strong classifiers determines final facial expression result. The proposed algorithm guarantees the realtime and enhanced accuracy by utilizing fastness and accuracy of Adaboost classification algorithm and reliability of neural network based recognition algorithm. In this paper, the proposed algorithm recognizes five facial expressions such as neutral, happiness, sadness, anger and surprise and achieves 86~95% of accuracy depending on the expression types in real time.

• The Journal of Korea Institute of Information, Electronics, and Communication Technology
• /
• v.8 no.6
• /
• pp.491-496
• /
• 2015

Cloud computing has emerged with promise to decrease the cost of server additional cost and expanding the data storage and ease for computer resource sharing and apply the new technologies. However, Cloud computing also raises many new security concerns due to the new structure of the cloud service models. Therefore, the secure user authentication is required when the user is using cloud computing. This paper, we propose the enhanced AdaBoost algorithm for access cloud security zone. The AdaBoost algorithm despite the disadvantage of not detect a face inclined at least 20, is widely used because of speed and responsibility. In the experimental results confirm that a face inclined at least 20 degrees tilted face was recognized. Using the FEI Face Database that can be used in research to obtain a result of 98% success rate of the algorithm perform. The 2% failed rate is due to eye detection error which is the people wearing glasses in the picture.

• Journal of KIISE
• /
• v.42 no.9
• /
• pp.1124-1132
• /
• 2015

In order to train a supervised machine learning algorithm, it is necessary to have non-biased labels and a sufficient amount of training data. However, it is difficult to collect the required non-biased labels and a sufficient amount of training data to develop an automatic English Composition scoring system. In addition, an English writing assessment is carried out using a multi-faceted evaluation of the overall level of the answer. Therefore, it is difficult to choose an appropriate machine learning algorithm for such work. In this paper, we show that it is possible to alleviate these problems through ensemble learning. The results of the experiment indicate that the ensemble technique exhibited an overall performance that was better than that of other algorithms.