• Journal of IKEEE
• /
• v.18 no.4
• /
• pp.581-585
• /
• 2014

In this paper, we propose a pedestrian detection by applying the HOG feature using ROI. Conventional HOG method has high accuracy, but shows the disadvantage of slow processing speed. By applying the ROI to the conventional method reduce computations for unnecessary area. Therefore proposed method improves the processing speed. In order to set the ROI area, we propose a structure that combined odd frames and even frames. Odd frame is in charge of operation for the entire area. And even frame does the operation for the ROI area. Implementation results of proposed method maintaining the same accuracy as the conventional method show a 20% improved performance of 8.3 frames per second.

• Proceedings of the KIEE Conference
• /
• 2015.07a
• /
• pp.1345-1346
• /
• 2015

본 논문에서는 보행자 및 배경 이미지로부터 HOG-PCA 특징을 추출하고 다항식 기반 RBFNNs(Radial Basis Function Neural Network) 패턴분류기과 최적화 알고리즘을 이용하여 보행자를 검출하는 시스템 설계를 제안한다. 입력 영상으로부터 보행자를 검출하기 위해 전처리 과정에서 HOG(Histogram of oriented gradient) 알고리즘을 통해 특징을 추출한다. 추출된 특징은 고차원이므로 패턴분류기 분류 시 많은 연산과 처리속도가 따른다. 이를 개선하고자 PCA (Principal Components Analysis)을 사용하여 저차원으로의 차원 축소한다. 본 논문에서 제안하는 분류기는 pRBFNNs 패턴분류기의 효율적인 학습을 위해 최적화 알고리즘인 PSO(Particle Swarm Optimization)을 사용하여 구조 및 파라미터를 최적화시켜 모델의 성능을 향상시킨다. 사용된 데이터로는 보행자 검출에 널리 사용되는 INRIA2005_person data set에서 보행자와 배경 영상을 각각 1200장을 학습 데이터, 검증 데이터로 구성하여 분류기를 설계하고 테스트 이미지를 설계된 최적의 분류기를 이용하여 보행자를 검출하고 검출률을 확인한다.

• Proceedings of the Korean Information Science Society Conference
• /
• 2006.10b
• /
• pp.505-508
• /
• 2006

홍채인식 시스템은 홍채영역 검출, 홍채특징 코드 생성, 그리고 홍채코드 비교 판단의 과정으로 이루어져 있다. 기존의 논문이나 연구들의 대부분은 앞에서 나열한 홍채인식 시스템의 과정의 일부만을 수정하여 성능개선, 즉 인식속도 향상과 인식률 향상 등을 꾀하였다. 이에 반해, 본 논문에서는 홍채인식 과정 전체의 개선을 통하여, 획기적으로 홍채인식 시간을 단축시키는 홍채인식 방법을 제안 하였다. Hough Transform과 Vertical & Horizontal Histogram을 사용한 홍채영역 검출, gradient를 사용한 홍채코드 생성, 그리고 variance를 이용하는 홍채코드의 비교와 판단 과정을 빠르고 단순한 알고리즘으로 구성하여, 홍채인식 속도를 개선하였다. 본 논문에서 제안한 홍채인식 시스템의 성능을 실험한 결과, mobile 환경에서 실시간으로 사용 할 수 있는 속도와 기존 홍채인식 시스템과 비슷한 홍채인식률을 나타내었다.

• International Journal of Contents
• /
• v.11 no.3
• /
• pp.47-53
• /
• 2015

Visual object tracking is a fundamental problem in the field of computer vision, as it needs a proper model to account for drastic appearance changes that are caused by shape, textural, and illumination variations. In this paper, we propose a feature-based visual-object-tracking method with a sparse representation. Generally, most appearance-based models use the gray-scale pixel values of the input image, but this might be insufficient for a description of the target object under a variety of conditions. To obtain the proper information regarding the target object, the following combination of features has been exploited as a corresponding representation: First, the features of the target templates are extracted by using the HOG (histogram of gradient) and LBPs (local binary patterns); secondly, a feature-based sparsity is attained by solving the minimization problems, whereby the target object is represented by the selection of the minimum reconstruction error. The strengths of both features are exploited to enhance the overall performance of the tracker; furthermore, the proposed method is integrated with the particle-filter framework and achieves a promising result in terms of challenging tracking videos.

• International Journal of Contents
• /
• v.13 no.4
• /
• pp.12-15
• /
• 2017

Among the possible stereotactic body radiation therapy (SBRT) modalities used to treat patients with metastatic spinal tumors, this study compared Cyberknife, tomotherapy, and volumetric modulated arc radiotherapy (VMAT). We established treatment plans for each of them modality and quantitatively analyzed the dose evaluation factors of the dose-volume histogram (DVH) for all spinal bones, focusing on the tumor and spinal cord, in order to examine the usefulness of VMAT. For the treatment planning dose, the mean dose ($D_{max}$) and $D_{5%}$ showed statistical differences in the target dose, but no difference was shown in the spinal cord dose. For the DVH indices, tomotherapy showed the best performance was the best in terms of uniformity index, while VMAT showed better performance was better than the other two modalities in terms of the conformity index and the dose gradient index. VMAT had a much shorter treatment time than Cyberknife and tomotherapy. These findings suggest that VMAT FFF is the most effective therapy for SBRT of patients with metastatic spinal tumors for whom a high dose of radiation is prescribed.

• The Journal of Korean Society for Radiation Therapy
• /
• v.19 no.2
• /
• pp.113-122
• /
• 2007

Purpose: We proposed the method using dose-volume Histogram index to compare prospective plan trials in tomotherapy planning optimization. Materials and Methods: For 3 patients in cranial region, thorax and abdominal region, we acquired computed tomography images with PQ 5000 in each case. Then we delineated target structure and normal organ contour with pinnacle Ver 7.6c, after transferred each data to tomotherapy planning system (hi-art system Ver 2.0), we optimized 3 plan trials in each case that used differ from beam width, pitch, importance. We analyzed 3 plan trials in each region with isodose distribution, dose-volume histogram and dose statistics. Also we verified 3 plan trials with specialized DVH-indexes that is dose homogeneity index in target organ, conformity index around target structure and dose gradient index in non-target structures. Results: We compared with the similarity of results that the one is decide the best plan trial using isodose distribution, dose volume histogram and dose statistics, and the another is using DVH-indexes. They all decided the same plan trial to better result in each case. Conclusion: In some of case, it was appeared a little difference of results that used to DVH-index for comparison of plan trial in tomotherapy by special goal in it. But because DVH-index represented both dose distribution in target structure and high dose risk about normal tissue, it will be reasonable method for comparison of many plan trials before the tomotherapy treatments.

• The Journal of Korean Society for Radiation Therapy
• /
• v.19 no.2
• /
• pp.91-97
• /
• 2007

Purpose: To evaluate the feasibility of a commercial ion chamber array for intensity modulated radiation therapy (IMRT) quality assurance (QA) was performed IMRT patient-specific QA Materials and Methods: A use of IMRT patient-specific QA was examined for nasopharyngeal patient by using 6MV photon beams. The MatriXX (Wellhofer Dosimetrie, Germany) was used for IMRT QA. The case of nasopharyngeal cancer was performed inverse treatment planning. A hybrid dose distribution made on the CT data of MatriXX and solid phantom all of the same gantry angle (0$^\circ$). The measurement was acquired with geometrical condition that equal to hybrid treatment planning. The $\gamma$-index (dose difference 3%, DTA 3 mm) histogram was used for quantitative analysis of dose discrepancies. An absolute dose was compared at the high dose low gradient region. Results: The dose distribution was shown a good agreement by gamma evaluation. A proportion of acceptance criteria was 95.8%, 97.52%, 96.28%, 98.20%, 97.78%, 96.64% and 92.70% for gantry angles were 0$^\circ$, 55$^\circ$, 110$^\circ$, 140$^\circ$, 220$^\circ$, 250$^\circ$ and 305$^\circ$, respectively. The absolute dose in high dose low gradient region was shown reasonable agreement with the RTP calculation within $\pm$3%. Conclusion: The MatriXX offers the dosimetric characteristics required for performing both relative and absolute measurements. If MatriXX use in the clinic, it could be simplified and reduced the IMRT patient-specific QA workload. Therefore, the MatriXX is evaluated as a reliable and convenient dosimeter for IMRT patient-specific QA.

• Journal of Institute of Control, Robotics and Systems
• /
• v.16 no.12
• /
• pp.1150-1158
• /
• 2010

In this study, the Polynomial-based Radial Basis Function Neural Networks is proposed as one of the recognition part of overall face recognition system that consists of two parts such as the preprocessing part and recognition part. The design methodology and procedure of the proposed pRBFNNs are presented to obtain the solution to high-dimensional pattern recognition problem. First, in preprocessing part, we use a CCD camera to obtain a picture frame in real-time. By using histogram equalization method, we can partially enhance the distorted image influenced by natural as well as artificial illumination. We use an AdaBoost algorithm proposed by Viola and Jones, which is exploited for the detection of facial image area between face and non-facial image area. As the feature extraction algorithm, PCA method is used. In this study, the PCA method, which is a feature extraction algorithm, is used to carry out the dimension reduction of facial image area formed by high-dimensional information. Secondly, we use pRBFNNs to identify the ID by recognizing unique pattern of each person. The proposed pRBFNNs architecture consists of three functional modules such as the condition part, the conclusion part, and the inference part as fuzzy rules formed in 'If-then' format. In the condition part of fuzzy rules, input space is partitioned with Fuzzy C-Means clustering. In the conclusion part of rules, the connection weight of pRBFNNs is represented as three kinds of polynomials such as constant, linear, and quadratic. Coefficients of connection weight identified with back-propagation using gradient descent method. The output of pRBFNNs model is obtained by fuzzy inference method in the inference part of fuzzy rules. The essential design parameters (including learning rate, momentum coefficient and fuzzification coefficient) of the networks are optimized by means of the Particle Swarm Optimization. The proposed pRBFNNs are applied to real-time face recognition system and then demonstrated from the viewpoint of output performance and recognition rate.

• Journal of Korea Multimedia Society
• /
• v.16 no.2
• /
• pp.169-179
• /
• 2013

Change detection algorithms take two image frames and return the locations of newly introduced objects which cause differences between the images. This paper presents a new change detection method, which classifies intensity changes due to introduced objects, reflected light and shadow from the objects to their neighborhood, and the noise, and exactly localizes the introduced objects. For classification and localization, first we analyze the histogram of the intensity difference between two images, and estimate multiple threshold values. Second we estimate candidate object boundaries using the gradient difference between two images. Using those threshold values and candidate object boundaries, we segment the frame difference image into multiple regions. Finally we classify whether each region belongs to the introduced objects or not using textures in the region. Experiments show that the proposed method exactly localizes the objects in various scenes with different lighting.

• Journal of KIISE
• /
• v.44 no.1
• /
• pp.44-50
• /
• 2017

Pedestrian detection, an important component of autonomous driving and driving assistant system, has been extensively studied for many years. In particular, image based pedestrian detection methods such as Hierarchical classifier or HOG and, deep models such as ConvNet are well studied. The evaluation score has increased by the various methods. However, pedestrian detection requires high sensitivity to errors, since small error can lead to life or death problems. Consequently, further reduction in pedestrian detection error rate of autonomous systems is required. We proposed a new method to detect pedestrians and reduce the error rate by using the Faster R-CNN with new developed pedestrian training data sets. Finally, we compared the proposed method with the previous models, in order to show the improvement of our method.