• Proceedings of the Korea Society for Industrial Systems Conference
• /
• 2002.06a
• /
• pp.323-329
• /
• 2002

멀티미디어 정보검색 중 내용 기반 영상검색은 색상, 질감, 형태 등의 영상 내용 특징들을 이용하여 검색하는 방법으로, 색상과 질감 특징을 이용한 검색 시스템이 일반적으로 널리 소개되고 있다. 그러나 형태가 서로 다른 영상에서는 색상과 질감 특징에 의한 검색 방법은 유사 영상검색에서 오류를 수반할 수 있다. 그래서 본 논문에서는 영상의 윤곽선 에 의한 전역 형태 특징으로 허용 가능한 범주 이내로 유사도 영상을 필터링한 후 형태정보의 히스토그램을 이용하여 유사도 검색을 함으로써 정확도를 놀일 수 있는 시스템을 개발한다.

• Proceedings of the Korean Information Science Society Conference
• /
• 2003.04a
• /
• pp.713-715
• /
• 2003

특성 추출은dimensionality reduction technique로서 잡음을 제거하기 위해 사용되는 중요한 전처리 방식이다. 이러한 과정을 통해 데이터의 크기를 줄일 수 있으며 학습의 정확성 및 이해도를 높일 수 있다. Classification에 사용되는 다양한 특성 추출방식들이 존재하는 반면에 클러스터링에 적용될 수 있는 방식들은 양적으로도 많이 부족하며 존재하는 방식들도 대부분 사용되는 클러스터링 알고리즘 자체에 의존적인 실세계 어플리케이션에는 적용하기 부적합한 Wrapper 방식을 도입하고 있다. 본 논문에서는 클러스터링 알고리즘으로부터 독립적인 필터 솔루션(filter solution)을 제안하였다. 이 방식은 클러스터를 가진 데이터와 가지지 않고 있는 데이터 사이의 point-to-point 거리 히스토그램의 차이에 기반하고 있다.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2018.06a
• /
• pp.345-346
• /
• 2018

차종 별 교통량 자료는 도로의 유지관리나 분석 등의 행정 처리 업무에 필요한 기본 자료임과 동시에 각종 연구에 활용된다. 본 시스템은 그 일환으로서 화물차나 일반차량을 구분하여 특정 도로의 화물차 비율이나 교통량을 파악하는데 활용할 수 있다. 머신 러닝 알고리즘 중에서 높은 성능을 보이는 Support Vector Machine (SVM) 알고리즘을 이용하여 도로 위의 일반차량과 화물차를 구분하였다. 우선, 화물차와 일반차량의 차이를 구분하고자 각각의 영상에 대해 Histogram of Oriented Gradients (HOG) 기반 특징점을 추출하고 이에 따라 1 차원 벡터로 표현된 데이터를 SVM 으로 분류하여 구분한다.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2023.11a
• /
• pp.741-744
• /
• 2023

장면 전환 검출에서 단일 특성을 사용하는 경우 발생 가능한 정확도 감소의 문제를 해결하기 위해 색상 히스토그램 분포 차 분석과 특징점 추출 알고리즘을 활용한 방법을 제안한다.

• Journal of Internet Computing and Services
• /
• v.20 no.4
• /
• pp.91-102
• /
• 2019

In autonomous driving systems, the ability to classify pedestrians in images captured by cameras is very important for pedestrian safety. In the past, after extracting features of pedestrians with HOG(Histogram of Oriented Gradients) or SIFT(Scale-Invariant Feature Transform), people classified them using SVM(Support Vector Machine). However, extracting pedestrian characteristics in such a handcrafted manner has many limitations. Therefore, this paper proposes a method to classify pedestrians reliably and effectively using CNN's(Convolutional Neural Network) deep features and transfer learning. We have experimented with both the fixed feature extractor and the fine-tuning methods, which are two representative transfer learning techniques. Particularly, in the fine-tuning method, we have added a new scheme, called M-Fine(Modified Fine-tuning), which divideslayers into transferred parts and non-transferred parts in three different sizes, and adjusts weights only for layers belonging to non-transferred parts. Experiments on INRIA Person data set with five CNN models(VGGNet, DenseNet, Inception V3, Xception, and MobileNet) showed that CNN's deep features perform better than handcrafted features such as HOG and SIFT, and that the accuracy of Xception (threshold = 0.5) isthe highest at 99.61%. MobileNet, which achieved similar performance to Xception and learned 80% fewer parameters, was the best in terms of efficiency. Among the three transfer learning schemes tested above, the performance of the fine-tuning method was the best. The performance of the M-Fine method was comparable to or slightly lower than that of the fine-tuningmethod, but higher than that of the fixed feature extractor method.

• The Journal of The Korea Institute of Intelligent Transport Systems
• /
• v.19 no.3
• /
• pp.129-136
• /
• 2020

In this paper, we propose a thermal imagery-based object detection algorithm for low-light level nighttime surveillance system. Many features selected by Haar-like feature selection algorithm and existing Adaboost algorithm are often vulnerable to noise and problems with similar or overlapping feature set for learning samples. It also removes noise from the feature set from the surveillance image of the low-light night environment, and implements it using the lightweight extended Haar feature and adaboost learning algorithm to enable fast and efficient real-time feature selection. Experiments use extended Haar feature points to recognize non-predictive objects with motion in nighttime low-light environments. The Adaboost learning algorithm with video frame 800*600 thermal image as input is implemented with CUDA 9.0 platform for simulation. As a result, the results of object detection confirmed that the success rate was about 90% or more, and the processing speed was about 30% faster than the computational results obtained through histogram equalization operations in general images.

• Journal of Broadcast Engineering
• /
• v.9 no.1
• /
• pp.2-7
• /
• 2004

In this paper, we proposed the content-based music information retrieval technique using some MPEG-7 low-level descriptors. Especially, pitch information and timbral features can be applied in music genre classification, music retrieval, or QBH(Query By Humming) because these can be modeling the stochasticpattern or timbral information of music signal. In this work, we restricted the music domain as O.S.T of movie or soap opera to apply broadcasting system. That is, the user can retrievalthe information of the unknown music using only an audio clip with a few seconds extracted from video content when background music sound greeted user's ear. We proposed the audio feature set organized by MPEG-7 descriptors and distance function by vector distance or ratio computation. Thus, we observed that the feature set organized by pitch information is superior to timbral spectral feature set and IFCR(Intra-Feature Component Ratio) is better than ED(Euclidean Distance) as a vector distance function. To evaluate music recognition, k-NN is used as a classifier

• Journal of Korea Multimedia Society
• /
• v.15 no.12
• /
• pp.1442-1448
• /
• 2012

The problem of Metal Area Segmentation (MAS) in X-ray CT images is a very hard task because of metal artifacts. This research features a practical yet effective method for MAS in X-ray CT images that exploits both projection image and reconstructed image spaces. We employ the Relevant Neighbor Area (RNA) idea [1] originally developed for projection image inpainting in order to create a novel feature in the projection image space that distinctively represents metal and near-metal pixels with opposite signs. In the reconstructed result of the feature image, application of a simple thresholding technique provides accurate segmentation of metal areas due to nice separation of near-metal areas from metal areas in its histogram.

• 제어로봇시스템학회:학술대회논문집
• /
• 2003.10a
• /
• pp.1782-1787
• /
• 2003

A face recognition is currently the field which many research have been processed actively. But many problems must be solved the previous problem. First, We must recognize the face of the object taking a location various lighting change and change of the camera into account. In this paper, we proposed that new method to find feature within fast and correct computation time after scanning PC camera and ID card picture. It converted RGB color space to YUV. A face skin color extracts which equalize a histogram of Y ingredient without the Luminance. After, the method use V' ingredient which transforms V ingredient of YUV and then find the face feature. The result of the experiment shows getting correct input face image from ID Card picture and camera.

• Communications for Statistical Applications and Methods
• /
• v.10 no.2
• /
• pp.457-469
• /
• 2003

In this paper, we have proposed the efficient algorithm that can segment the video scene change using a various statistical characteristics obtained from by applying the wavelet transformation for each frames. Our method firstly extracts the histogram features from low frequency subband of wavelet-transformed image and then uses these features to detect the abrupt scene change. Second, it extracts the edge information from applying the mesh method to the high frequency subband of transformed image. We quantify the extracted edge information as the values of variance characteristic of each pixel and use these values to detect the gradual scene change. And we have also proposed an algorithm how extract the proper key frame from segmented video scene. Experiment results show that the proposed method is both very efficient algorithm in segmenting video frames and also is to become the appropriate key frame extraction method.