Earticle

본 논문에서는 로봇에서 감정을 표현하기 위하여 사용하는 동작을 쉽게 생성하기 위한 알고리즘을 제 시하였다. 이와 같은 목적을 위하여, 키 프레임 방식을 사용하였으며, 각각의 키 프레임은 감정을 표현하 기 위한 로봇의 핵심 포즈이다. 감정을 표현하기 위한 동작은 여러 개의 키 프레임으로 이루어지며, 동작 의 개선을 위하여 키 프레임의 삽입, 또는 삭제가 가능하다. 이와 같이 여러 개의 키 프레임으로 구성되 는 동작을 부드럽게 이루어지도록하기 위하여, 기존의 PWM 방식 모터 제어 방법을 개선하여 적용하였 다. 키 프레임 시퀀스를 편집하기 위한 프로그래밍 시스템도 개발되었다.

The purpose of this study was to develop the algorithm to create a new predetermined gesture for the emotion expression easily. For the purpose of easy creation of predetermined gesture, key frame concept was adopted. The key frame which was used in this study is a robot pose that is critical in order to express the emotion. A predetermined gesture is composed of several key frame. And one or more key frame can be added or removed easily to modify the gesture. For the smooth movement of the robot, modified PWM is proposed. A programming system was also developed for the editing of the key frame sequence and simulation of the servo motor movement.

To manipulate large video contents, effective video indexing and retrieval are required. A large number of video indexing and retrieval algorithms have been presented for frame-wise user query or video content query whereas a relatively few video sequence matching algorithms have been proposed for video sequence query. In this paper, we propose an efficient algorithm that extracts key frames using color histograms and matches the video sequences using edge features. To effectively match video sequences with a low computational load, we make use of the key frames extracted by the cumulative measure and the distance between key frames, and compare two sets of key frames using the modified Hausdorff distance. Experimental results with real sequence show that the proposed video sequence matching algorithm using edge features yields the higher accuracy and performance than conventional methods such as histogram difference, Euclidean metric, Battachaya distance, and directed divergence methods.

본 연구는 뇌파측정기술(EEG)의 ERP와 판별분석을 이용하여 이용자 기반의 비디오 키프레임들을 자동으로 추출할 수 있는 방안을 제안하였다. 구체적으로, 20명의 피험자를 대상으로 한 실험을 통해서 이미지 처리 과정을 다섯 개의 정보처리단계들(자극주목, 자극지각, 기억인출, 자극/기억 대조, 적합 판정)로 구분하여 각 단계별로 적합한 뇌파측정기술의 ERP 유형을 제안하여 검증해 보았다. 검증 결과, 각 단계별로 서로 다른 ERP 유형(N100, P200, N400, P3b, P600)을 나타냈다. 또한 세 그룹(적합, 부분적합 및 비적합 프레임)간을 구별할 수 있는 중요한 변수들로는 P3b에서 P7의 양전위 최고값과 FP2의 음전위 최저값의 잠재기로 나타났고, 이러한 변수들을 이용해 판별분석을 수행한 후 적합 및 비적합 프레임들을 분류할 수 있었다.

This study proposed a key-frame automatic extraction method for video storyboard surrogates based on users' cognitive responses, EEG signals and discriminant analysis. Using twenty participants, we examined which ERP pattern is suitable for each step, assuming that there are five image recognition and process steps (stimuli attention, stimuli perception, memory retrieval, stimuli/memory comparison, relevance judgement). As a result, we found that each step has a suitable ERP pattern, such as N100, P200, N400, P3b, and P600. Moreover, we also found that the peak amplitude of left parietal lobe (P7) and the latency of FP2 are important variables in distinguishing among relevant, partial, and non-relevant frames. Using these variables, we conducted a discriminant analysis to classify between relevant and non-relevant frames.

본 연구는 뇌파측정기술(EEG)과 사건관련유발전위 P3b를 활용하여 이용자의 인지적 반응을 측정한 후 비디오 키프레임을 자동으로 추출할 수 있는지의 가능성을 조사해 보았다. 20명의 피험자들을 대상으로 뇌파를 측정하고 분석한 결과, 적합 이미지 자극 시 좌측 두정엽 영역이 우측 두정엽 영역보다 더 활성화되며, 좌우측간 두정엽 영역의 활성화 정도가 유의한 차이를 보였다. 비적합 이미지 자극 시에는 좌측 두정엽 영역이 적합 이미지보다 덜 활성화되고, 두정엽 영역의 좌우간 활성화도 유의한 차이가 없는 것으로 나타났다. 이외에, 모든 채널의 평균값(MGFP1)의 잠재기, 채널 동시성 패턴 등에서도 두 자극간에 차이를 보여 뇌파측정기술에 기반한 키프레임 자동 추출이 가능한 것으로 확인되었다.

This study examined the feasibility of using EEG signals and ERP P3b for extracting video key-frames based on users' cognitive responses. Twenty participants were used to collect EEG signals. This research found that the average amplitude of right parietal lobe is higher than that of left parietal lobe when relevant images were shown to participants; there is a significant difference between the average amplitudes of both parietal lobes. On the other hand, the average amplitude of left parietal lobe in the case of non-relevant images is lower than that in the case of relevant images. Moreover, there is no significant difference between the average amplitudes of both parietal lobes in the case of non-relevant images. Additionally, the latency of MGFP1 and channel coherence can be also used as criteria to extract key-frames.

In the current Internet environment, a lot of multimedia information is used; such as text, sound, image, and video. Among those information video is the most valuable and meaningful data. Therefore, the key frames of a video are very important role to recognize the whole sequence contents. In this paper, we propose a new key frame extraction scheme using chi-square and color histogram. The proposed key frame extraction scheme showed the better results compared to the previous studies. The proposed scheme results will be applied to the many computer game programs for the fast scene detection and the recognition of the game contents.

온라인 게임을 비롯한 많은 인터넷 매체 중에 사회적 통제를 요구하는 유해한 동영상이 아무른 재제 없이 무분별하게 배포됨으로서 심각한 역기능이 발생되고 있다. 이는 청소년의 정신 건강에 큰 문제를 발생하게 되면 사회적, 제도적 규제가 절실하다. 본 논문에서는 영상이 가지는 컬러 공간 정보를 이용하여 동영상의 유해성을 판별할 수 있는 대표 프레임 선정 기법을 제안한다. 이 기법을 기존의 히스토그램 정보를 이용한 유해한 대표 프레임 선정 기법과 비교하여 매우 효과적임을 증명하였다. 본 연구 결과는 유해 동영상을 판별하는 도구에 적용하면 매우 효과적일 것이다.

In this paper, proposes the technique which selects the Key-frame that use color space information. And prove is very effective comparing with technique that use histogram information. The result of this paper when uses in the tool which discriminates the noxiousness of the video is effective.

In order to reconstruct a 3D model in video sequences, to select key frames that are easy to estimate a geometric model is essential. This paper proposes a method to easily extract informative frames from a handheld video. The method combines selection criteria based on appropriate-baseline determination between frames, frame jumping for fast searching in the video, geometric robust information criterion (GRIC) scores for the frame-to-frame homography and fundamental matrix, and blurry-frame removal. Through experiments with videos taken in indoor space, the proposed method shows creating a more robust 3D point cloud than existing methods, even in the presence of motion blur and degenerate motions.

Key frame extraction is considered as one of the most critical issues in content-based video retrieval technology (CBVR). In this paper, an efficient key frame extraction algorithm based on unsupervised clustering and compressive sensing is proposed. Firstly, three types of filters with various scales are employed to generate high dimensional feature of each frame in one shot, which will be projected to low dimensional feature by a very sparse random projection matrix that satisfies the condition of Restricted Isometry Property (RIP), and then sub-shot segmentation is conducted by an unsupervised clustering method in order to divide one shot into sub-shot collections, in which each class of clustering represents one sub-shot. Finally, the Bhattacharyya coefficient is used to measure the similarity between frame and class center, the frame with the maximum similarity value is selected as the key frame in each sub-shot. The experimental results demonstrate that the proposed method could extract key frames efficiently and properly.

Aiming at changing high computational complexity, underdeveloped real time, low recognition rate of dynamic gesture recognition algorithms, this paper present a real-time dynamic gesture trajectory recognition method based on key frame extraction and HMM. Key frames are selected without keeping track of all the details of one dynamic gesture, which is based on difference degree between frames. The trajectory data stream, sorted by the time-warping algorithm, is used to construct the Hidden Markov Method model of dynamic gesture. Finally, optimal transition probabilities are employed to implement dynamic gesture recognition. The result of this experiment implies that this method has high robustness and real time. The average recognition rate of dynamic gesture (0~9) is up to 87.67%, and average time efficiency is 0.46s.

We propose efficient video retrieval scheme which use fast key frame extraction in DCT domain. Our scheme extract key frame using the edge histogram difference which is extracted in compressed domain for I-frames. And the video retrieval is implemented using Hausdorff distance function about edge histogram of key frame. This approach enables fast content-based video retrieval of the compressed video content without decompression process. Experimental results show our scheme is very fast and efficient.

WMV(Winoow Media Video)는 ASF(Advanced Streaming Format)파일 포맷을 기반으로 하는 마이크로 소프트사의 대표적인 스트리밍 파일 포맷이다. WMV 파일 포맷은 MPEG-4를 기반으로 내용 기반 부호화를 통해 영상을 객체 단위로 나누어 처리하므로 사용자의 의도에 따라 다양한 조작과 디스플레이가 가능하다. 또한 영상 데이터를 패킷에 붙여 네트워크를 통해 스트리밍 서비스도 가능하다. 따라서 본 논문에서는 WMV파일 포맷 분석을 통해 영상의 가장 중요한 Hey-Frame을 추출과 데이터 헤더 부분을 제외한 실제 영상 스트림의 시작 위치를 찾는 알고리즘을 설계 및 구현하였다. 이로써 실제 스트리ald 서비스외에도 다양한 멀티디미어 서비스에 적용 가능한 방법을 마련하였다.

In this paper, we propose a new and fast method for assigning the number of key frames to each shot. At first we segment the entire video sequence into elementary content unit called shots and then the key frame allocation is performed by calculating the accumulated value of AF(activity function). The proposed algorithm is based on the amount of content variation using DC images extracted from compressed video. By assigning the number of key frames to the shot that has the largest value of content function, one key frame is assigned at a time until you run out of given all key frames. The main advantage of our proposed method is that we do not need to use time-exhaustive computations in allocating the key frames over the shot and can perform it fully automatically.

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.

본 논문에서는 동영상 데이터베이스에서 Key-frame을 검색하는 방법을 제안한다. 본 논문에서는 Key-frame을 검색하기위해 컬러 피쳐를 공간영역에서 추출하지 않고 wavelet transform 영역에서 컬러 피쳐를 추출하는 방법을 제안한다. wavelet transform 의 저주파 밴드는 영상전체의 특징을 잘 나타내고 고주파 밴드는 texture 와 국부적인 컬러 특성을 잘 나타낸다. 색인과정 알고리즘은 영상의 크기를 정규화하고 RGB 컬러공간에서 HSV 컬러 공간으로 변환을 하여, H, S, V 각 채널에 대해 Daubechies' wavelet transform을 수행한 후 변환 영역에서 피쳐를 추출하게 된다. 색인을 위한 피쳐로 wavelet 계수와 lowest 밴드의 평균과 표준편차를 추출하였다. 효율적인 검색을 위해 검색은 2단계로 수행된다. 먼저 평균과 표준편차만을 이용한 1차 검색을 통해 2차 검색의 후보 영상들을 추출하고 2차 검색에서는 1차 검색 통과 영상들에 대해서만 wavelet 계수들을 비교하여 최종 검색 결과를 얻게 된다. 검색결과 기존의 컬러 피쳐를 이용한 방법보다 우수한 검색결과를 얻을 수 있었다.

영상을 구성하는 프레임 중에 키프레임은 일반적으로 영상 정보를 효과적으로 요약하거나 용이한 분석을 위해 선정된다. 화상이 가진 의미는 인물/사물 등의 객체탐지를 통해 추출되는데, 기존의 키프레임 관련 연구는 영상이 가지는 의미를 반영하는 키프레임을 찾아내기 어렵다. 본 논문에서는 영상이 가지는 특정 의미가 있다고 할 때 이를 반영하는 키프레임을 효과적으로 추출하는 방법을 실험적으로 탐구하였다. 구체적으로 영상을 통할하는 의미를 피로라고 가정하고 영상의 졸음 인식 관련 연구에 사용되는 DDD 데이터셋을 이용하여 효과적인 키프레임 추출 기법을 적용해 보았으며, 실험 결과 졸음이라는 특정 정보에 대한 해석을 도울 수 있는 의미 있는 요약을 제공하는 키프레임들을 효과적으로 추출하는 분석 기법을 찾아낼 수 있었다.

Due to the popularity of mobile phones, Peer-to-Peer (P2P) content sharing in mobile phones is highly desirable, especially in a mobile data connection. In addition, there has been an increasing demand for content sharing applications for P2P communication in Android-based smartphones as ubiquitous mobile devices. However, due to wireless network limits, unstable characteristic and restricted storage space, mobile users confront challenges in establishing connections and sharing video frames with other users. To solve this issue, we propose an architectural design and implementation of key frame sharing application that uses Chord, a distributed lookup protocol to establish a content sharing platform for mobile devices, combined with the Chord SDK based on Android. Our proposed system enables users easily to share key information on video data in real-time over a wireless network without manual processing of devices which join or leave the group and cost associated with the 3G/4G network.

This paper presents an approach for dynamic hand gesture recognition by using algorithm based on 3D Convolutional Neural Network (3D_CNN), which is later extended to 3D Residual Networks (3D_ResNet), and the neural network based key frame selection. Typically, 3D deep neural network is used to classify gestures from the input of image frames, randomly sampled from a video data. In this work, to improve the classification performance, we employ key frames which represent the overall video, as the input of the classification network. The key frames are extracted by SegNet instead of conventional clustering algorithms for video summarization (VSUMM) which require heavy computation. By using a deep neural network, key frame selection can be performed in a real-time system. Experiments are conducted using 3D convolutional kernels such as 3D_CNN, Inflated 3D_CNN (I3D) and 3D_ResNet for gesture classification. Our algorithm achieved up to 97.8% of classification accuracy on the Cambridge gesture dataset. The experimental results show that the proposed approach is efficient and outperforms existing methods.

In this paper, we proposed a mechanism that prevents indiscreet use of digital contents, verifies created content's copyrights to provide services to granted user and protects digital contents by law by authenticating the original digital content whenever an infringement of copyright occurs in UCC environments. The proposed mechanism uses specific information and features of contents as copyrights authentication information without additional information. Also, provides the fact of violation by inferring the modification of the original digital contents. That means this mechanism infers same or similar value from the contents; fraudulent use of content, modification of content color, modification of content format, modification of content resolution and illegal use of frame not principal key frame. In that point, we found that the inferred value from the content differed according to features of content modification.

본 논문에서는 MPEG 비디오 스트림을 분석하여 DCT DC 계수를 추출하고 이들로 구성된 DC 이미지로부터 제안하는 robust feature를 이용하여 shot 검출을 수행한 후 각 feature들의 통계적 특성을 이용하여 스트림의 특징에 따라 weight를 부가하여 구해진 characterizing value의 시간 변화량을 구한다. 추해진 변화량의 local maxima와 local minima는 비디오 스트림에서 각각 가장 특징적인 frame과 평균적인 frame을 나타낸다. 이 순간의 shot을 구함으로서 효과적이고 빠른 시간 내에 key frame을 추출한다. 추출되어진 key frame에 대하여 원영상을 복원한 후, 색인을 위하여 다수의 parameter를 구하고, 사용자가 질의한 영상에 대해서 이들 파라메터를 구하여 key frame들과 가장 유사한 대표영상들을 검색한다. 실험결과 일반적인 방법보다 더 나은 결과를 보였고, 높은 검색율을 보였다.

This Paper suggests a new feature for shot detection, using the proposed robust feature from the DC image constructed by DCT DC coefficients in the MPEG video stream, and proposes the characterizing value that reflects the characteristic of kind of video (movie, drama, news, music video etc.). The key frames are pulled out from many frames by using the local minima and maxima of differential of the value. After original frame(not do image) are reconstructed for key frame, indexing process is performed through computing parameters. Key frames that are similar to user's query image are retrieved through computing parameters. It is proved that the proposed methods are better than conventional method from experiments. The retrieval accuracy rate is so high in experiments.