• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2020.07a
• /
• pp.727-731
• /
• 2020

인공신경망의 예측 결과에 대한 원인을 분석하는 것은 모델을 신뢰하기 위해 필요한 작업이다. 이에 컴퓨터 비전 분야에서는 돌출맵 또는 히트맵의 형태로 모델이 어떤 내용을 근거로 예측했는지 시각화 하는 모델 해석 방법들이 제안되었다. 하지만 오디오 분야에서는 스펙트로그램 상의 시각적 해석이 직관적이지 않으며, 실제 어떤 소리를 근거로 판단했는지 이해하기 어렵다. 따라서 본 연구에서는 히트맵의 청취 분석 시스템을 제안하고, 이를 활용한 음향 장면 분류 모델의 히트맵 청취 분석 실험을 진행하여 인공신경망의 예측 결과에 대해 사람이 이해할 수 있는 설명을 제공할 수 있는지 확인한다.

• Proceedings of the Korea Society of Environmental Biology Conference
• /
• 2004.05a
• /
• pp.57-57
• /
• 2004

• Communications of the Korean Institute of Information Scientists and Engineers
• /
• v.11 no.5
• /
• pp.41-50
• /
• 1993

• Proceedings of the Korea Contents Association Conference
• /
• 2007.11a
• /
• pp.698-702
• /
• 2007

Recently, it isn't hard to find split screen effects in films, TV dramas and commercial films which composes a screen with different frames of images. The TV series '24' which was produced in USA in the year 2001 and is expecting the release of its 7th series next year frequently exhibits such split screen methods. '24' is a TV drama of the events taking place in one day which is produced in the form of real-time mode and consists of 24 episodes which is an hour long. Using split screen method, the drama effectively delivers the events simultaneously taking place at different locations. Moreover, the divided frames of each screen relates to one another by means of collision, synthesis and ect. which captures the vision of the spectators. This study aims at analysing the relationships between the frames which consists split screen scenes in the drama '24' and discover its eye capturing attractions using film montage theories.

• Journal of KIISE:Software and Applications
• /
• v.34 no.6
• /
• pp.549-561
• /
• 2007

Vision-based scene understanding is to infer and interpret the context of a scene based on the evidences by analyzing the images. A probabilistic approach using Bayesian networks is actively researched, which is favorable for modeling and inferencing cause-and-effects. However, it is difficult to gather meaningful evidences sufficiently and design the model by human because the real situations are dynamic and uncertain. In this paper, we propose a learning method of Bayesian network that reduces the computational complexity and enhances the accuracy by searching an efficient BN structure in spite of insufficient evidences and training data. This method represents the domain knowledge as ontology and builds an efficient hierarchical BN structure under constraint rules that come from the ontology. To evaluate the proposed method, we have collected 90 images in nine types of circumstances. The result of experiments indicates that the proposed method shows good performance in the uncertain environment in spite of few evidences and it takes less time to learn.

• The Journal of the Korea Contents Association
• /
• v.9 no.6
• /
• pp.92-103
• /
• 2009

With the help of computer graphics technologies, the visual effects techniques using these technologies replaced most of special effects techniques which had been used for early films. For these changes, directors and visual effects creators make an effect in a scene through their mutual agreement in contemporary films. However, they undergo a lot of trial-and-error while making a visual effects scene because they cannot perfectly communicate their ideas due to the director's narrative language, and also because of the visual effect creator's language of computer graphics technology. This research suggests the design of a visual effects data retrieval system for efficient communication between directors and visual effects creators. This application provides the means to search a database analyzing visual effects scenes extracted from 14 remarkable movies in visual effect history by narrative and visual effects technique. They can search visual effects scenes using this application. also, this data can foster communication with directors and creators so they can make an efficient production pipeline.

• Journal of the Korea Society of Computer and Information
• /
• v.11 no.3
• /
• pp.193-201
• /
• 2006

This paper presents a method that allows for detection of all rapid and gradual scene changes. The method features a combination of the current color histogram and the local $X^{2}-test$. For the purpose of this paper, the $X^{2}-test$ scheme outperforming existing histogram-based algorithms was transformed, and a local $X^{2}-test$ in which weights were applied in accordance with the degree of brightness was used to increase detection efficiency in the segmentation of color values. This Method allows for analysis and segmentation of complex time-varying images in the most general and standardized manner possible Experiments were performed to compare the proposed local $X^{2}-test$ method with the current $X^{2}-test$ method.

• Journal of the Korea Society of Computer and Information
• /
• v.26 no.8
• /
• pp.31-37
• /
• 2021

In this paper, we propose a method to effectively search by automatically classifying scenes in which specific images such as pitching or swing appear in baseball game images using a CNN(Convolution Neural Network) model. In addition, we propose a video scene search system that links the classification results of specific motions and game records. In order to test the efficiency of the proposed system, an experiment was conducted to classify the Korean professional baseball game videos from 2018 to 2019 by specific scenes. In an experiment to classify pitching scenes in baseball game images, the accuracy was about 90% for each game. And in the video scene search experiment linking the game record by extracting the scoreboard included in the game video, the accuracy was about 80% for each game. It is expected that the results of this study can be used effectively to establish strategies for improving performance by systematically analyzing past game images in Korean professional baseball games.

• The Journal of the Korea Contents Association
• /
• v.8 no.7
• /
• pp.26-34
• /
• 2008

Due to spread of movies or online games which are applied with computer animation techniques, we can easily see scenes where numerous characters appear. In the case of large-scale crowd animation, if one were to increase reality of the scene, features of system would be lowered, and if one were to increase functioning of system, reality of the scene would be lowered. In realizing large-scale crowd animation with seafloor environment as background, the paper analyzed and applied elements that affect behavioral types of fishes; and by using concept of crowd, the paper enabled each group or object to control their behavioral type; by comparing and contrasting real-time calculation method as calculation method for animation and hybrid calculation method which is mixed calculation method, the paper seeks to find a method that increases functioning of the system while also expresses natural scenes.

• Journal of the Korea Convergence Society
• /
• v.9 no.11
• /
• pp.217-225
• /
• 2018

Different eye-heights due to individuals' body heights may cause different sensory experiences against the same visual scene, eventually leading to their longer-term psycho-social and developmental individual differences. Accordingly, the present study compared sensory properties of photographs for the same scene taken from two different camera-heights (i.e., eye-heights). Two sets of photographs were taken in parallel from two cameras attached to a different height on the same pedestrian's body. Analysis of the photographs revealed that both the levels of visual saliency and complexity were greater for the photographs taken from the high eye-height than those from the low eye-height. The results indicate a possible difference in sensory properties of visual scenes perceived from two different heights, potentially exposing taller individuals to richer and more diverse sensory experiences than shorter individuals.