• Journal of Internet Computing and Services
• /
• v.11 no.1
• /
• pp.85-102
• /
• 2010

In this paper, we propose a novel method of analyzing video and representing the relationship among characters based on their contexts in the video sequences, namely Character-Net. As a huge amount of video contents is generated even in a single day, the searching and summarizing technologies of the contents have also been issued. Thereby, a number of researches have been proposed related to extracting semantic information of video or scenes. Generally stories of video, such as TV serial or commercial movies, are made progress with characters. Accordingly, the relationship between the characters and their contexts should be identified to summarize video. To deal with these issues, we propose Character-Net supporting the extraction of major characters in video. We first identify characters appeared in a group of video shots and subsequently extract the speaker and listeners in the shots. Finally, the characters are represented by a form of a network with graphs presenting the relationship among them. We present empirical experiments to demonstrate Character-Net and evaluate performance of extracting major characters.

• Proceedings of the Korean Society of Computer Information Conference
• /
• 2013.01a
• /
• pp.29-30
• /
• 2013

본 논문에서는 영화나 소설과 같은 콘텐츠의 스토리를 가시화해서 보여주는 시스템에 대해 제안하고 설명한다. 스토리를 가시화 해주기 위해 등장인물들 간의 관계를 모형화하는 Character-net 방법론을 채용하였고 스토리 진행에 따른 Character-net 변화를 분석하여 보여주는 시스템을 개발하였다. 시스템은 Character-net 변화 실행창과 등장인물 중심성 시계열 그래프 창으로 구성하였다. 두 개 창을 통해 스토리 차원의 검색이 가능토록 하였다. 본 논문에서는 스토리 가시화 시스템에 대해 설명하고 추가적으로 필요한 사항들에 대해 논의한다.

• Journal of Broadcast Engineering
• /
• v.23 no.1
• /
• pp.53-62
• /
• 2018

It is necessary to analyze the character that are a key element of the story in order to analyze the story. Current character analysis methods such as Character-net and RoleNet are not sufficient to classify the roles of supporting characters by only analyzing the results of the final accumulated stories. It is necessary to study the time series analysis method according to the story progress in order to analyze the role of supporting characters rather than the accumulated story analysis method. In this paper, we propose a method to classify helpers as a mentor and a best friend through 3-D visualization of Character-net and evaluate the accuracy of the method. WebGL is used to configure the interface for 3D visualization so that anyone can see the results on the web browser. It is also proposed that rules to distinguish mentors and best friends and evaluated their performance. The results of the evaluation of 10 characters selected for 7 films confirms that they are 90% accurate.

• Journal of the Korea Society of Computer and Information
• /
• v.14 no.10
• /
• pp.241-249
• /
• 2009

Recently, a number of researches related to video annotation and representation have been proposed to analyze video for searching and abstraction. In this paper, we have presented a method to provide the picture elements of conversational participants in video and the enhanced representation of the characters using those elements, collectively called Character-net. Because conversational participants are decided as characters detected in a script holding time, the previous Character-net suffers serious limitation that some listeners could not be detected as the participants. The participants who complete the story in video are very important factor to understand the context of the conversation. The picture elements for detecting the conversational participants consist of six elements as follows: subtitle, scene, the order of appearance, characters' eyes, patterns, and lip motion. In this paper, we present how to use those elements for detecting conversational participants and how to improve the representation of the Character-net. We can detect the conversational participants accurately when the proposed elements combine together and satisfy the special conditions. The experimental evaluation shows that the proposed method brings significant advantages in terms of both improving the detection of the conversational participants and enhancing the representation of Character-net.

• Journal of Information Processing Systems
• /
• v.14 no.1
• /
• pp.205-217
• /
• 2018

Deep learning using convolutional neural networks (CNNs) is being studied in various fields of image recognition and these studies show excellent performance. In this paper, we compare the performance of CNN architectures, KCR-AlexNet and KCR-GoogLeNet. The experimental data used in this paper is obtained from PHD08, a large-scale Korean character database. It has 2,187 samples of each Korean character with 2,350 Korean character classes for a total of 5,139,450 data samples. In the training results, KCR-AlexNet showed an accuracy of over 98% for the top-1 test and KCR-GoogLeNet showed an accuracy of over 99% for the top-1 test after the final training iteration. We made an additional Korean character dataset with fonts that were not in PHD08 to compare the classification success rate with commercial optical character recognition (OCR) programs and ensure the objectivity of the experiment. While the commercial OCR programs showed 66.95% to 83.16% classification success rates, KCR-AlexNet and KCR-GoogLeNet showed average classification success rates of 90.12% and 89.14%, respectively, which are higher than the commercial OCR programs' rates. Considering the time factor, KCR-AlexNet was faster than KCR-GoogLeNet when they were trained using PHD08; otherwise, KCR-GoogLeNet had a faster classification speed.

• Journal of the Korea Society of Computer and Information
• /
• v.18 no.3
• /
• pp.119-126
• /
• 2013

In this paper, we propose and describe the system to visualize the story of contents such as movies and novels. Character-net is applied as story model in order to visualize story. However, it is the form to be accumulated for total movie story, though it can depict the relationship between characters. We have developed the system that analyzes and shows the variation of Character-net and characters' tendency in order to represent story variation depending on movie progression. This system is composed by two windows that can play and analyze sequential Character-nets by time, and can analyze time variant graph of characters' degree centrality. First window has a function that supports to find important story points like the scenes that main characters appear or meet firstly. Second window supports a function that track each character's tendency or a variation of his tendency through analyzing in-degree graph and out-degree. This paper describes the proposed system and discusses additional requirements.

• Journal of the Korea Society of Computer and Information
• /
• v.14 no.11
• /
• pp.169-178
• /
• 2009

There are various approaches that retrieve information from video. However, previous approaches have considered just object information and relationship between objects without story information to retrieve contents. To retrieve exact information at video, we need analyzing approach based on characters and community since these are body of story proceeding. Therefore, this paper describes video information retrieval methodology based on character information. Characters progress story to form relationship through conversations. We can analyze the relationship between characters in a story with the methods that classifies role grades and clusters communities of characters. In this paper, for these, we propose the Character-net and describe how to classify role grades and cluster communities at Character-net. And we show this method to be efficient.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.20 no.9
• /
• pp.1657-1665
• /
• 2016

Various fields are being researched through Deep Learning using CNN(Convolutional Neural Network) and these researches show excellent performance in the image recognition. In this paper, we provide streamlined GoogLeNet of CNN architecture that is capable of learning a large-scale Korean character database. The experimental data used in this paper is PHD08 that is the large-scale of Korean character database. PHD08 has 2,187 samples for each character and there are 2,350 Korean characters that make total 5,139,450 sample data. As a training result, streamlined GoogLeNet showed over 99% of test accuracy at PHD08. Also, we made additional Korean character data that have fonts that are not in the PHD08 in order to ensure objectivity and we compared the performance of classification between streamlined GoogLeNet and other OCR programs. While other OCR programs showed a classification success rate of 66.95% to 83.16%, streamlined GoogLeNet showed 89.14% of the classification success rate that is higher than other OCR program's rate.

• Smart Media Journal
• /
• v.9 no.2
• /
• pp.92-98
• /
• 2020

This paper proposes a vehicle license plate recognition system using light weight deep learning models without high-end server. The proposed license plate recognition system consists of 3 steps: [license plate detection]-[character area segmentation]-[character recognition]. SSD-Mobilenet was used for license plate detection, ResNet with localization was used for character area segmentation, ResNet was used for character recognition. Experiemnts using Samsung Galaxy S7 and LG Q9, accuracy showed 85.3% accuracy and around 1.1 second running time.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2016.05a
• /
• pp.97-100
• /
• 2016

Using the CNN(Convolutinal Neural Network), Deep Learning for variety of fields are being developed and these are showing significantly high level of performance at image recognition field. In this paper, we show the test accuracy which is learned by large-scale training data, over 5,000,000 of Korean characters. The architecture of CNN used in this paper is KCR(Korean Character Recognition)-AlexNet newly created based on AlexNet. KCR-AlexNet finally showed over 98% of test accuracy. The experimental data used in this paper is large-scale Korean character database PHD08 which has 2,187 samples for each Korean character and there are 2,350 Korean characters that makes total 5,139,450 sample data. Through this study, we show the excellence of architecture of KCR-AlexNet for learning PHD08.