• 제어로봇시스템학회논문지
• /
• 제19권11호
• /
• pp.1017-1021
• /
• 2013

The traffic signal has been widely used in the transport system with a fixed time interval currently. This kind of setting time was determined based on experience for vehicles to generate a waiting time while allowing pedestrians crossing the street. However, this strict setting causes inefficient problems in terms of economic and safety crossing. In this research, we propose a monitoring algorithm to detect, track and check pedestrian crossing the crosswalk by the patterns of behavior. This monitoring system ensures the safety for pedestrian and keeps the traffic flow in efficient. In this algorithm, pedestrians are detected by using HOG feature which is robust to illumination changes in outdoor environment. According to a complex computation, the parallel process with the GPU as well as CPU is adopted for real-time processing. Therefore, pedestrians are tracked by the relationship of hue channel in image sequence according to the predefined pedestrian zone. Finally, the system checks the pedestrians' crossing on the crosswalk by its HOG based behavior patterns. In experiments, the parallel processing by both GPU and CPU was performed so that the result reaches 16 FPS (Frame Per Second). The accuracy of detection and tracking was 93.7% and 91.2%, respectively.

• 한국지능시스템학회논문지
• /
• 제15권7호
• /
• pp.813-816
• /
• 2005

귀 영역 검출은 무구속 귀 인식 시스템에서 중요한 요소 중에 하나이다. 본 논문에서는 얼굴 측면 영상에서의 귀 영역 검출방법을 제안한다. 제안하는 방법은 모양정보와 색상정보를 활용하는 인간의 인식과정을 모방하여 만들었다. 먼저 획득된 영상에서 피부색을 이용하여 얼굴영역을 검출하고, 검출된 얼굴영역에서 에지정보를 이용하여 귀후보 영역을 검출한다. 그리고 실제 귀영역인지를 검증하기 위해서는 통계적 학습 이론에 근거한 SVM(Support Vector Machine)을 이용한다 제안된 방법은 조명이 안정적인 실내 환경에서 높은 검출율을 보였다.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제10권11호
• /
• pp.5605-5623
• /
• 2016

We propose a face recognition method that utilizes the LCP face descriptor. The proposed method applies a LoG mask to extract a face contour response, and employs the LCP algorithm to produce a binary pattern representation that ensures high recognition performance even under the changes in illumination, noise, and aging. The proposed LCP algorithm produces excellent noise reduction and efficiency in removing unnecessary information from the face by extracting a face contour response using the LoG mask, whose behavior is similar to the human eye. Majority of reported algorithms search for face contour response information. On the other hand, our proposed LCP algorithm produces results expressing major facial information by applying the threshold to the search area with only 8 bits. However, the LCP algorithm produces results that express major facial information with only 8-bits by applying a threshold value to the search area. Therefore, compared to previous approaches, the LCP algorithm maintains a consistent accuracy under varying circumstances, and produces a high face recognition rate with a relatively small feature vector. The test results indicate that the LCP algorithm produces a higher facial recognition rate than the rate of human visual's recognition capability, and outperforms the existing methods.

• International Journal of Fuzzy Logic and Intelligent Systems
• /
• 제14권4호
• /
• pp.332-339
• /
• 2014

Face classification has wide applications in security and surveillance. However, this technique presents various challenges caused by pose, illumination, and expression changes. Face recognition with long-distance images involves additional challenges, owing to focusing problems and motion blurring. Multiple frames under varying spatial or temporal settings can acquire additional information, which can be used to achieve improved classification performance. This study investigates the effectiveness of multi-frame decision-level fusion with photon-counting linear discriminant analysis. Multiple frames generate multiple scores for each class. The fusion process comprises three stages: score normalization, score validation, and score combination. Candidate scores are selected during the score validation process, after the scores are normalized. The score validation process removes bad scores that can degrade the final output. The selected candidate scores are combined using one of the following fusion rules: maximum, averaging, and majority voting. Degraded facial images are employed to demonstrate the robustness of multi-frame decision-level fusion in harsh environments. Out-of-focus and motion blurring point-spread functions are applied to the test images, to simulate long-distance acquisition. Experimental results with three facial data sets indicate the efficiency of the proposed decision-level fusion scheme.

• 제어로봇시스템학회논문지
• /
• 제11권5호
• /
• pp.423-428
• /
• 2005

This paper proposes a new edge detection method using a $3{\times}3$ ideal binary pattern and lookup table (LUT) for the mobile robot localization without any parameter adjustments. We take the mean of the pixels within the $3{\times}3$ block as a threshold by which the pixels are divided into two groups. The edge magnitude and orientation are calculated by taking the difference of average intensities of the two groups and by searching directional code in the LUT, respectively. And also the input image is not only partitioned into multiple groups according to their intensity similarities by the histogram, but also the threshold of each group is determined by fuzzy reasoning automatically. Finally, the edges are determined through non-maximum suppression using edge confidence measure and edge linking. Applying this edge detection method to the mobile robot localization using projective invariance of the cross ratio. we demonstrate the robustness of the proposed method to the illumination changes in a corridor environment.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제11권8호
• /
• pp.3965-3983
• /
• 2017

We propose a novel eye tracking method that can estimate 3D world coordinates using an infrared (IR) stereo camera for indoor and outdoor environments. This method first detects dark evidences such as eyes, eyebrows and mouths by fast multi-level thresholding. Among these evidences, eye pair evidences are detected by evidential reasoning and geometrical rules. For robust accuracy, two classifiers based on multiple layer perceptron (MLP) using gradient local binary patterns (GLBPs) verify whether the detected evidences are real eye pairs or not. Finally, the 3D world coordinates of detected eyes are calculated by region-based stereo matching. Compared with other eye detection methods, the proposed method can detect the eyes of people wearing sunglasses due to the use of the IR spectrum. Especially, when people are in dark environments such as driving at nighttime, driving in an indoor carpark, or passing through a tunnel, human eyes can be robustly detected because we use active IR illuminators. In the experimental results, it is shown that the proposed method can detect eye pairs with high performance in real-time under variable illumination conditions. Therefore, the proposed method can contribute to human-computer interactions (HCIs) and intelligent transportation systems (ITSs) applications such as gaze tracking, windshield head-up display and drowsiness detection.

• Journal of Multimedia Information System
• /
• 제5권2호
• /
• pp.105-110
• /
• 2018

Autonomous cars require the integration of multiple communication systems for driving safety. Many carmakers unveil mirrorless concept cars aiming to replace rear and sideview mirrors in vehicles with camera monitoring systems, which eliminate blind spots and reduce risk. This paper presents optical vehicle-to-vehicle (V2V) communications for autonomous mirrorless cars. The flicker-free light emitting diode (LED) light sources, providing illumination and data transmission simultaneously, and a high speed camera are used as transmitters and a receiver in the OCC link, respectively. The rear side vehicle transmits both future action data and vehicle type data using a headlamp or daytime running light, and the front vehicle can receive OCC data from the camera that replaces side mirrors so as not to prevent accidents while driving. Experimental results showed that action and vehicle type information were sent by LED light sources successfully to the front vehicle's camera via the OCC link and proved that OCC-based V2V communications for mirrorless cars can be a viable solution to improve driving safety.

• 한국지능시스템학회논문지
• /
• 제18권3호
• /
• pp.422-428
• /
• 2008

본 논문에서는 spatial gradient를 이용한 강인한 물체 추출 방법을 제안한다. 제안한 방법은 먼저 복잡한 환경과 다양한 빛의 변화에 의해 나타나는 에러 값 등을 해결하기 위해 기존에 제안된 입력 영상과 기준 영상에서 밝기와 색 성분을 이용하여 최초 배경을 제거한다. 배경을 제거한 다음, 그림자로 인식되어 전경 영역에 추가된 부분을 RGB 칼라 모델과 정규화 된 RGB 칼라 모델을 이용하여 제거하고, HSI 칼라 모델을 이용하여 불필요한 정보 값을 갖는 영역을 제거한다. 마지막으로, 배경으로 인식되어 전경으로부터 제거된 부분을 입력 영상의 공간상 정보인 spatial gradient와 HSI 칼라 모델을 이용하여 복구하는 방법을 제안한다. 마지막으로, 본 논문에서 제안한 알고리즘은 복잡하고 다양한 실내 외 환경에서의 실험을 통해 그 응용 가능성을 증명한다.

• 한국산업정보학회논문지
• /
• 제24권2호
• /
• pp.47-55
• /
• 2019

본 논문에서는 조명의 영향을 받지 않는 깊이 정보를 이용한 번호판 검출 방법을 제안한다. 깊이 정보를 통해 블록 내 화소들의 3차원 카메라 좌표를 구하고, 이를 통해 블록 내 평면의 인자를 계산한다. 그 후 인접한 블록간의 평면의 법선 벡터들을 비교하여 유사도를 측정한다. 평면 유사도가 높을 경우 두 블록이 한 평면에 속해 있다고 간주하여 그룹화함으로써 평면 영역을 검출한다. 검출된 평면 영역에 대해 깊이 정보를 이용하여 영역의 높이와 너비를 실제 번호판과 비교하여 번호판을 검출한다.

• 대한전자공학회:학술대회논문집
• /
• 대한전자공학회 2005년도 추계종합학술대회
• /
• pp.553-556
• /
• 2005

We present a vision-based hand gesture recognition system for object manipulation in virtual space. Most conventional hand gesture recognition systems utilize a simpler method for hand detection such as background subtractions with assumed static observation conditions and those methods are not robust against camera motions, illumination changes, and so on. Therefore, we propose a statistical method to recognize and detect hand regions in images using geometrical structures. Also, Our hand tracking system employs multiple cameras to reduce occlusion problems and non-synchronous multiple observations enhance system scalability. Experimental results show the effectiveness of our method.