Scene Change Detection Research Articles

지역적 정보를 이용한 장면 전환 검출

This paper proposes a Scene Change Detection method using the local decision tree and clustering. The local decision tree detects cluster boundaries wherein local scenes occur, in such a way as to compare time similarity distributions among the difference values between detected scenes and their adjacent frames, and group an unbroken sequence of frames with similarities in difference value into a cluster unit.

Enhanced sensory processing accompanies successful detection of change for real-world sounds

Change deafness is the inability of listeners to detect changes occurring in their auditory environment. It is a matter of some debate whether change deafness occurs because of a failure of auditory-specific processes or a failure of more general semantic/verbal memory. To address this issue, we measured event-related potentials (ERPs) to pairs of scenes consisting of naturalistic auditory objects while listeners made a same/different judgment for scenes presented before and after an interruption. ERPs to the post-change scene revealed an enhanced early sensory response (N1) and an enhanced late positivity (P3) for detected changes. Change detection performance was better when there was a large acoustic spread among the objects within Scenes 1 and 2, suggesting that the deficits reflected by the ERP components during change deafness are related to successfully segregating the pre- and post-change objects. We also found that a separate sensory response (P2) reflects implicit, unconscious change detection. Overall, the results provide evidence that auditory-specific sensory processing is critical for both explicit and implicit change detection in natural auditory scenes.

Scene Change Detection Using Multiple Histograms for Motion-Compensated Frame Rate Up-Conversion

In this paper, we propose a new scene change detection method using multiple histograms. The proposed method produces multiple histograms of split blocks for consecutive frames. Then, it computes the optimal threshold value using automatic thresholding based on the Otsu method and decides whether the scene change occurs or not using the difference between threshold values in successive frames. In the experiments for the subjective evaluation, the proposed method correctly identified the scene change, thereby preventing interpolated frames with poor image quality due to block artifacts. In the objective evaluation using the F1 score, the proposed method improved the accuracy of the scene change detection by up to 0.461 compared with the benchmark methods.

The Timing of Change Detection and Change Perception in Complex Acoustic Scenes

We investigated how listeners perceive the temporal relationship of a light flash and a complex acoustic signal. The stimulus mimics ubiquitous events in busy scenes which are manifested as a change in the pattern of on-going fluctuation. Detecting pattern emergence inherently requires integration over time; resulting in such events being detected later than when they occurred. How does delayed detection time affect the perception of such events relative to other events in the scene? To model these situations, we use rapid sequences of tone pips with a time-frequency pattern that changes from random to regular (“REG-RAND”) or vice versa (“RAND-REG”). REG-RAND transitions are detected rapidly, but RAND-REG take longer to detect (∼880 ms post nominal transition). Using a Temporal Order Judgment task, we instructed subjects to indicate whether the flash appeared before or after the acoustic transition. The point of subjective simultaneity between the flash and RAND-REG does not occur at the point of detection (∼880 ms post nominal transition) but ∼470 ms closer to the nominal acoustic transition. In a second experiment we halved the tone pip duration. The resulting pattern of performance was qualitatively similar to that in Experiment 1, but scaled by half. Our results indicates that the brain possesses mechanisms that survey the proximal history of an on-going stimulus and automatically adjust perception so as to compensate for prolonged detection time, thus producing more accurate representations of scene dynamics. However, this readjustment is not complete.

동영상 카투닝 시스템을 위한 자동 프레임 추출 기법

멀티미디어 산업의 발달과 함께 디지털 콘텐츠 시장의 확산을 가져오고 있다. 그 중 인터넷 만화와 같은 디지털 만화 시장의 확장은 급속하게 커지고 있어서, 콘텐츠의 부족과 다양성 때문에 동영상 카투닝에 대한 연구가 계속되고 있다. 지금까지는 동영상 카투닝은 비사실적 렌더링과 말풍선에 초점이 맞추어졌으나, 이러한 것들을 적용하기 위해서는 카투닝 서비스에 적합한 프레임 추출이 우선시 되어야만 한다. 기존의 방법으로는 동영상의 장면전환이 일어나는 샷(shot)안의 프레임을 추출하여, 사용자가 지정한 영역을 임의의 색상으로 렌더링(Rendering)하는 시스템이 있다. 하지만 이러한 방법은 사람의 손을 거치는 반자동적인 방법으로서 정확한 프레임 추출을 위해 사람의 손을 거쳐야하는 단점이 있다. 따라서 본 논문에서는 이러한 문제점을 해결하고, 보다 정확한 카투닝에 적용할 프레임을 추출하기 위해 오디오 및 비디오 분리를 통한 방법을 제안한다. 먼저 동영상으로부터 오디오와 비디오를 분리한다. 오디오는 먼저 MFCC와 영교차율의 특징을 추출하고, 이 특징 정보를 미리 학습된 데이터와 GMM 분류기를 통하여 음악, 음성, 음악+음성으로 분류한 후 음성 영역을 설정한다. 비디오는 히스토그램을 이용한 방법과 같은 일반적인 장면전환 프레임을 추출 후 얼굴 검색을 통해서 만화에서 의미가 있는 프레임을 추출한다. 그 후 음성 영역내에 얼굴이 존재하는 장면전환 프레임이나 일정 시간동안 음성이 지속되는 영역 중 장면전환 프레임을 추출하여 동영상 카투닝에 적합한 프레임을 자동으로 추출한다. While the broadband multimedia technologies have been developing, the commercial market of digital contents has also been widely spreading. Most of all, digital cartoon market like internet cartoon has been rapidly large so video cartooning continuously has been researched because of lack and variety of cartoon. Until now, video cartooning system has been focused in non-photorealistic rendering and word balloon. But the meaningful frame extraction must take priority for cartooning system when applying in service. In this paper, we propose new automatic frame extraction method for video cartooning system. At frist, we separate video and audio from movie and extract features parameter like MFCC and ZCR from audio data. Audio signal is classified to speech, music and speech+music comparing with already trained audio data using GMM distributor. So we can set speech area. In the video case, we extract frame using general scene change detection method like histogram method and extract meaningful frames in the cartoon using face detection among the already extracted frames. After that, first of all existent face within speech area image transition frame extract automatically. Suitable frame about movie cartooning automatically extract that extraction image transition frame at continuable period of time domain.

Scene Change Detection Algorithm for H.264 using Dynamic Threshold and Fast Motion Estimation Algorithm based on PSO

Most of the existed scene change detection algorithms use fixed thresholds for finding the scene change. These thresholds are obtained by empirically or they must be calculated before the detection after the whole sequence is obtained. If Videos having high scene complexity and variation, then the Performance of most scene change algorithms decreases considerably. In this paper, we study the correlation between local statistical characteristics, scene duration and scene change. Based on this analysis, we further propose and implement a scene change algorithm for H.264 codec, defining an automated, dynamic threshold model with fast motion estimation algorithm with low complexity which can efficiently trace scene changes.

X2히스토그램을 이용한 수면의 질 발전

Sleep is very important physiology to our human, about one third of human life was sent over to sleep. This paper measures of sleep and proposes sleep quality and future direction in order to improve the sleep environment. Sleep measure was determined by using X2 histogram that is one of the scene change detection method. X2 histogram method is one of the statistical scene change detection and is used in many studies because of the histogram method performs better than the other. And find out their relationship by entering the degree of fatigue, alcohol, and hungry in order to develop quality of sleep and extracting to tossing and turning according to each situation.

Robust video super resolution algorithm using measurement validation method and scene change detection

Abstract Explicit motion estimation is considered a major factor in the performance of classical motion-based super resolution (SR) algorithms. To reconstruct video frames sequentially, we applied a dynamic SR algorithm based on the Kalman recursive estimator. Our approach includes a novel measurement validation process to attain robust image reconstruction results under inexplicit motion estimation. In our method, the suitability for high-resolution pixel estimation is determined by the accuracy of motion estimation. We measured the accuracy of the image registration result using the Mahalanobis distance between the input low-resolution frame and the motion compensated high-resolution estimation. We also incorporate an effective scene change detection method dedicated to the proposed SR approach for minimizing erroneous results when abrupt scene changes occur in the video frames. According to the ratio of well-aligned pixels (i.e., motion is compensated accurately) to the total number of pixels, we are able to detect sudden changes of scene and context in the input video. Representative experiments on synthetic and real video data show robust performance of the proposed algorithm in terms of its reconstruction quality even with errors in the estimated motion.

컬러 히스토그램을 활용한 수면의 질 향상

본 논문에서는 수면을 취하는 침실의 수면 환경 데이터를 수집하고, 얻어진 조건 데이터들과 수면간의 관계를 분석한다. 또한 수면을 취하는 사람으로부터 장면 전환을 검출하여 신체의 상황과 수면 과정의 반응, 신체의 감각과 자극들을 제시한다. 영상의 장면 전한 검출은 컬러 히스토그램을 사용하였다. 컬러 히스토그램은 이전 프레임과 현재 프레임사이의 컬러 히스토그램 차이 값을 이용하는 방법이다. 그리고 피곤한 정도, 음주의 정도, 그리고 공복의 정도를 입력하여 각 상황별 뒤척임을 추출하도록 하였다. 시스템의 실험에는 온/습도 센서와 조도센서로 구성되어 있는 H-MOTE2420 센서를 이용하였다. 본 논문에서는 현대인들에게 규칙적이고 안락한 수면을 유도하기 위하여 최적의 수면 환경을 제공함으로써 수면의 질을 높일 수 있도록 한다. In this paper we collect data concerning sleep environments in a bedroom and analyze the relationship between the collected condition data and sleep. In addition, this paper detects scene changes from the subjects in a sleeping state and presents the physical conditions, reactions during sleep, and physical sensations and stimuli. To detect scene changes in image sequences, we used color histogram for the difference between the preceding frame and the current frame. In addition, to extract the tossing and turning for different situations, the subjects were instructed to enter the level of fatigue, the level of drinking, and the level of stomach emptiness. For the sleep experiment system, we used the H-MOTE2420 Sensor composed of temperature, humidity, and light sensors. This paper is intended to provide the best sleep environment that enhances sleep quality, thus inducing people today to get regular and comfortable sleep.

Efficient Scene Change Algorithm Based on Statistical Feature in the DCT Domain

In this paper, we propose an effective scene change detection algorithm directly in compressed domain. The proposed scene change algorithm(including abrupt change and ) test the n frame and n+1 frame through the extracting the feature of each frame. When extract the features of frame, two-dimensional statistical feature m1-s from DCT coefficients without its inverse transform was computed, Divide m1-s space into 42 unequal partitions (subspaces) and count the numbers pr within the 42 subspaces (entries) as the feature vector to judge the different of the two frames. locating scene changes is operated by comparison tests. In comparison with existing representative techniques, the experimental results show the superiority of the proposed method in terms of precision and processing speed.

Change Detection in Complex Scenes: Hemispheric Contribution and the Role of Perceptual and Semantic Factors

The perceptual salience and semantic relevance of objects for the meaning of a scene were evaluated with multiple criteria and then manipulated in a change-detection experiment that used an original combination of one-shot and tachistoscopic divided-visual-field paradigms to study behavioural hemispheric asymmetry. Coloured drawings that depicted meaningful situations were presented centrally and very briefly (120 ms) and only the changes were lateralised by adding an object in the right or in the left visual hemifield. High salience and high relevance improved both response times (RTs) and accuracy, although the overall contribution of salience was greater than that of relevance. Moreover, only for low-salience changes did relevance affect speed. RTs were shorter when a change occurred in the left visual hemifield, suggesting a right-hemisphere advantage for detection of visual change. Also, men responded faster than women. The theoretical and methodological implications are discussed.

Global motion–compensated preprocessing algorithm for block-based frame-rate conversion

We propose a global motion (GM) compensated preprocessing algorithm for block-based frame rate conversion (FRC). The proposed method estimates camera motions such as zooms or rotations between two input frames, and accordingly, produces nontranslational GM-free pictures by performing GM compensation with respect to the temporal location where the FRC method reconstructs an intermediate frame. To reduce the computational load, the proposed method first subsamples input images, and then block-wise motion estimation (ME) is performed. With the ME results, the proposed method detects scene changes and nontranslational GMs. This allows us to determine whether to proceed with the GM estimation and compensation processes. The geometric motion model is adopted to describe camera motions with four parameters, and these values are iteratively found on a motion vector (MV) field. Experimental results show that the proposed algorithm achieves significant performance improvements of subsequent FRC methods.

Visuospatial and object working memory in naturalistic scene change detection

Visuospatial and object working memory in naturalistic scene change detection

A content-specific attenuation of change blindness: Preferential attention to animate beings in natural scenes

Considerable changes to an image often go undetected by observers if performed during a brief visual disruption (e.g. a saccade or global mask). This change blindness however is attenuated for regions and objects independently rated as “centres of interest” (Rensink, O'Regan, & Clark, 1997). This suggests that the detection of scene changes is mediated by the subjective “interest” of image features and their consequent attraction of visual attention. To conclude that some objects are preferentially attended due to their greater perceived interest however begs the more fundamental question: Why are some objects more interesting—more attention commanding—than others? One possible dimension for the objective prediction of “interest” and visual attention is that between animate and inanimate objects. In this experiment it was hypothesized that animate objects, because of their evolutionarily-persistent biological importance and propensity to change meaningfully, should be attended more closely than inanimate objects. This hypothesis was tested in a change-detection paradigm wherein subjects viewed a series of complex, natural scenes. In each scene, an object was either alternately added and deleted, reversed in orientation, or left the same in each display of the image. Changes to animate objects (people and animals) were detected more quickly and more frequently than changes to inanimate objects (plants, small artifacts, and large artifacts) in the first experiment and replication. Further experiments revealed that the observed advantages for detecting changes to animate objects were not attributable to low-level visual features or a learned expectation of change. These findings suggest that viewer interest and visual attention to natural scenes—although appearing largely subjective and idiosyncratic—can be partially accounted for by objective dimensions such as that between animate and inanimate objects.

Change detection in normal, jumbled and inverted scenes

Change detection in normal, jumbled and inverted scenes

비디오에서 오정색과 오간색 식별

색이란 인간이 조형 활동을 처음 시작할 때부터 사용되어진 조형언어로서 인간이 눈으로 볼 수 있는 모든 현상 세계에 존재하는 것이다. 본 논문에서는 장면 전환 검출에서 추출된 대표 프레임을 대상으로 한국의 전통색 조화를 식별한다. 전통색은 오정색과 오간색으로 분류하여 조화를 이루는지 판별하도록 한다. 빨강, 파랑, 노랑, 검정, 흰색을 오정색이라 부르고 분홍, 하늘색, 보라, 유황색, 초록을오 간색이라 하여 식별한다. 먼저, Canny 알고리즘을 이용하여 경계선을 추출한다. 그리고 경계선을 중심으로 색들을 레이블링 하고 클러스터링 한다. 마지막으로, 전통 색채 조화 식별 방법을 이용하여 전통색을 식별한다. 본 논문에서 제시한 연구는 실험을 통하여 우수성이 입증 되었다. As color was used for formative language since a human activity was beginning, all the symptoms in the world that the human eye can see is present. In this paper, we identify Korea traditional color harmony for extracted key frames from scene change detection. Traditional color is classified as 5-Jung-color and 5-Kan-color, and determine whether to harmony. Red, blue, yellow, black, and white, called 5-Jung-color and pink, blue, purple, sulfur, and green, called the 5-Kan-color was identified. First, we extract edge using Canny algorithm. And, we are labeling and clustering colors around the edge. Finally, we identify the traditional color using identification method of traditional color harmony. The proposed study in this paper has been proven through experiments.

Noise-bound method for detecting shadow-free scene changes in image sequences

Many image processing applications are confounded by both sensor noise and cast shadows. All image sensors add noise to a captured image that can reduce algorithm sensitivity and performance, and global filters or fixed thresholds are often applied to limit their effects. Cast shadows can appear as scene changes and are difficult to adequately detect and remove from images and image sequences. We couple image-noise statistics with a dual-illumination shadow-detection algorithm to provide a novel color-based method for shadow-free scene-change detection whose performance is bound by metamerism and image noise, and has only one variable--the desired confidence interval for noise separation.

An Optimized Dynamic Scene Change Detection Algorithm for H.264/AVC Encoded Video Sequences

Scene change detection plays an important role in a number of video applications, including video indexing, semantic features extraction, and, in general, pre- and post-processing operations. This paper deals with the design and performance evaluation of a dynamic scene change detector optimized for H.264/AVC encoded video sequences. The detector is based on a dynamic threshold that adaptively tracks different features of the video sequence, to increase the whole scheme accuracy in correctly locating true scene changes. The solution has been tested on suitable video sequences resembling real-world videos thanks to a number of different motion features, and has provided good performance without requiring an increase in decoder complexity. This is a valuable issue, considering the possible application of the proposed algorithm in post-processing operations, such as error concealment for video decoding in typical error prone video transmission environments, such as wireless networks.

Change detection in JPEG 2000 compressed video

A method for detecting scene changes from JPEG 2000 compressed images is presented. Change information and localisation is derived from changes in the local data rate. They can be extracted by analysing the code streams only. The computationally expensive decoding of image data is eliminated.

Single-object-based segmentation and coding technique for video surveillance system

We present an adaptive object segmentation based on scene change detection techniques. First, the algorithm adaptively adjusts to skip the number of frames to cope with the amount of motion displacement to cover the entire object shape. Next, the spatial processing consists of noise removal and boundary smoothing techniques to discard background content as well as to remove the background noise, to obtain a segmented object. To evaluate the quality of segmentation, both of the standard benchmarks and camera imaging are employed. Results show that the proposed algorithm can achieve low error ratios under various standard sequence testing. The segmentation algorithm combines with MPEG-4 coding for camera imaging of video surveillance. This system can successfully demonstrate real-time video surveillance, as the frame rate is 10 to 30 using CPU-based software implementation.