Annual Review 2011 > Division of Information Systems

Multimedia Systems Laboratory

Noriaki Asada


Jung-pil Shin

Associate Professor

Hirohide Demura

Associate Professor

Naru Hirata

Assistant Professor

Ryu Keun Ho

Visiting Researcher

Recognition, analysis and expression of images and letters are studied in the multimedia systems laboratory.

In addition we devote positive cooperation in Japanese deep space exploration with Aizu Research Cluster for Space Science (ARC-Space), Research Center for Advanced Information Science and Technology (CAIST), and aim to be one of the kernel organizations in this research field near future.

The 3-D motion and the 3-D shapes of the object are analyzed, understood and expressed by computer from the images taken by multi TV camera or video camera in the study of computer vision and/or the 3-D stereovision. Our work has greatly contributed in Hayabusa project. The succession of landing on Itokawa by Hayabusa in November, 2005 owed to preciseness of 3D shape model of Itokawa by University of Aizu very much. And also development of HARMONICS (HAyabusa Remote MONItering & Commanding System) has contributed very much in operation. We are also contributing in the next Asteroid exploration project, Hayabusa2, where a lab member is leading it as a principal investigator (PI).

Three lab members are also energetically contributing in KAGUYA (SELENE) project, the Japanese lunar exploring mission and playing a major role in LISM, the remote sensing camera system. The major research targets are centered on lunar shape recognition, such as lunar terrain analyses, crater counting, understanding and detailed whole lunar geographic mapping. Furthermore, they are engaged in analyzing the lunar mineral composition using the category classification technique from color information of lunar images and investigating the future moon surface land usage.

We are considering almost all kind of computer method to contribute lunar exploration, such as automatic crater counting, operation plan and log managing system, integration of lunar topological map from both Terrain camera (TC) data and Laser Altimeter (LALT) data, and so on.

Furthermore, we are engaged also in space exploration programs in Japan including planetary and an asteroid exploration inquiry plan and contributing in research in lunar and planetary science.

In addition, we are supporting Japanese deep space exploration with our original conception and cooperating with both domestic and international organization, while the computer technology becomes necessary in this research field. In the on-line handwriting character recognition system, we are engaged in studying on-line stroke-order and stroke-number free character recognition, signature verification, effect of alcohol to signing, signature evaluation, Kanji (Chinese characters in Japanese) learning system, handwritten style font generation system (handwriting synthesis with one's style), oriental writing brush implementation (3D virtual brush), and 3D character recognition using Wii controller. And also collaborations with the Republic of Korea are implemented by this lab.

Laboratory members have participated actively in scientific meetings, both in Japan and abroad. They have presented and participated in seminars, and in presentation of scientific results in fully refereed publications.

Refereed Journal Papers

[asada-01:2011, naru-01:2011]

N. Hirata, W. Kawamae, T. A. Dang, K. Kitazato, H. Demura, and N. Asada. Archive and Analysis system for observation data of irregular-shaped small bodies on polygon shape models. Journal of Space Science Informatics Japan, 1:133-140, 2012.

Archives and analysis tools of exploration data are important for scientific research in planetary sciences. Although geographic coordinate systems of planetary bodies are fundamental to construct such systems, a spherical coordinate system is failed on irregular-shaped small bodies. Here we propose a novel concept to manage coordinates on the surface of irregular-shaped small bodies with polygon shape models. With this concept, we develop a data archive system providing a location-oriented search function, and a 3D-geographical information system (3D-GIS) for small bodies. We also test a method to map polygon meshes on a sphere with the spherical parameterization technique to make mapping data independent from a specific polygon shape model.


Wondu Chang and Jungpil Shin. A Statistical Handwriting Model for Style-Preserving and Variable Character Synthesis. International Journal of Document Analysis and Recognition (IJDAR), 15(1):1-19, March 2012.

Synthesizing handwritten-style characters is an interesting issue in today’s handwriting analysis field. The purpose of this study was to artificially generate training data, foster a deep understanding of human handwriting, and promote the use of the handwritten-style computer fonts, in which the individuality or variety of the synthesized characters is considered important. Research considering such two properties together, however, are very rare. In this paper, a handwriting model is proposed to synthesize various handwritten characters while preserving the writer’s individuality from a limited number of training data, using a statistical approach. The proposed model was verified in single- and multiple-stroke characters, such as Arabic numbers, small English letters, and Japanese Kanji letters. Synthesized characters were evaluated in three ways. First, they were analyzed visually using the selected samples, and the relationship between the training and synthesized characters was explained. Second, the personalities and varieties of all the data were evaluated using a conventional writer verification method. Third, a questionnaire was developed and administered to evaluate the subjective responses of the users regarding the personal styles of the synthesized characters. The results prove that the proposed model stably synthesizes personalized characters by being invariant to the number of training data, whereas the variety increases gradually as the data increase.


Yukari Sato, Jungpil Shin, and Hsien-Chou Liao. Random Font Generation in Personal Style. International Journal of Engineering and Industries (IJEI), 2(3):21-30, Sep 2011.

This article presents a new method for generating random handwriting fonts. Random fonts are rendered differently every time they are output onto the screen. When generating handwriting fonts, it is important to render the individual’s handwriting style as this can facilitate and improve communication. The purpose of this research is to develop a system that simulates an individual’s handwriting style using randomly generated fonts. This method utilizes three main steps: dynamic programming (DP) for matching the original data points of input strokes, a minimum spanning tree method, and random font generation through interpolation. DP matching improves the accuracy of the minimum spanning tree. The minimum spanning tree connects similar shapes in order to produce a new font shape had original features truly. The interpolation method allows a generated font shape to be similar to real handwriting characters but have a new shape. As a result, our system generates random fonts that retain an individual handwriting features, maintaining the wealth of font variation.


Jungpil Shin and Kohei Fujiwara. Japanese Waseda Shorthand Recognition. International Journal of Engineering and Industries (IJEI), 2(3):9-20, Sep 2011.

Handwritten shorthand is a very fast way to record speech, at the same speed as a person speaks. Shorthand is used to transcript speech at conferences, the Diet, etc. However, when compared to the speed of writing, it takes a long time to read and transform shorthand sentences into regular text. Computer technology that recognizes shorthand has been investigated as a way to solve this problem, although Japanese shorthand recognition has not been actively researched. In this paper, we research online character recognition for Waseda shorthand, which is a kind of Japanese handwritten shorthand. Our shorthand recognition method has two stages: a segmentation stage and a classification stage. In addition, we used the smooth junction detection method in the segmentation stage of our system. Smooth junctions are feature points that can be used to segment an input stroke. For smooth junction detection, we used quadratic differentials to detect flexion points as the smooth junctions. In our experiments, five writers wrote 523 words (1781 characters), and systems both with and without smooth junction detection attempted to recognize the writing. The results showed that the recognition rate when using smooth junction detection improved by up to 48.09detection method. In future work, we would like to decrease the rate of recognition errors in smooth junction detection by preprocessing the writing. In addition, we would like to develop a method of dynamically defining thresholds and adding new rules for dominant point detection.


Jungpil Shin and Hsien-Chou Liao. An Interactive Map Search System using Wavelet and Shape Contexts. Journal of Advanced Computational Intelligence and Intelligent Informatics (JACIII), 15(9):1256-1262, Sep 2011.

In this paper a new interactive map search system is presented using shape context and bipartite graph matching. Shape context is used for measuring shape similarity and the recovering of point correspondences. After the above information is generated from the shape context, bipartite graph matching is used to obtain the optimal correspondence between two shapes. Hierarchical description is also used to increase the recognition rate. Shape context is a method to treat shapes as a set of points and generate the histogram of the distribution of points. Wavelet analysis is used in hierarchical description. In order to shorten the calculation time, piecewise linear approximation is implemented as the feature extraction method. The system lists the six most similar shapes to hand-written input shapes from the reference shapes, i.e., Japan’s 47 prefectures. Comparison results of linear matching, dynamic programming (DP) matching, and shape context with bipartite graph matching indicate that the 1st place recognition rates are 82evaluation result of hierarchical description shows that hierarchical approximation can improve the recognition rate from 92.45 to 94.97deepest-4 depth. These results show that the proposed method is effective on fulfilling the interactive map search system.


Takanori Kuroiwa and Jungpil Shin. Discovery of Efficient Chinese Characters for Handwritten-Style Font Generation. International Journal of Digital Content Technology and its Applications (JDCTA), 5(12):1-12, Dec 2011.

This paper proposes a method for discovering efficient Chinese characters so as to generate handwritten-style fonts effectively. This method is composed of four steps: Selection of Essential Input Characters, Collection of Tentative Input Characters, Elimination of Nonessential Input Characters, and Iterative Improvement. First, the characters that are essential for generating all other characters are selected. This reduces the cost of finding efficient input characters. Then, tentative input characters are collected. After that, nonessential characters are eliminated from the set of tentative input characters. Finally, efficient input characters are determined by reducing the number of tentative input characters one by one on the basis of the iterative improvement method. With this procedure, all 2,965 Chinese characters were generated using 126 efficient input characters. The characters with relatively few strokes were selected as efficient input characters. As a result, the burden of writing characters for font generation was greatly reduced.


Hsien-Chou Liao, Chun-Yang Lu, and Jungpil Shin. The incorporation of GPS and IP Cameras for People Tracking. GPS Solutions (SCI), (doi:10.1007/s10291-011-0242-8):1-13, Oct 2011.

Using the global positioning system (GPS) for people tracking continues to get easier. A person can transmit his/her GPS location from the carried mobile devices. The location is usually displayed as a dot on a digital map. However, a dot on the map is insufficient to reveal the person’s actual situation, e.g., an accident being happening. If the GPS is incorporated with an IP (Internet Protocol) camera, the camera image is critical in revealing the person’s actual situation and to improve the above-mentioned insufficient information. We present an approach to facilitate such incorporation. The approach consists of three phases: locating, tracking and monitoring collision. When the GPS coordinates of a person are within the field-of-view (FOV) of a camera, the approach enters the locating phase. The GPS coordinates are transformed to specify a candidate area (CA) in the image. The update of GPS coordinates is used to filter those moving objects within the CA until only one remains. After the person is located, he is being tracked using the shortest Euclidean distance method to find the most likely object in the next image. If the person collides with other objects while being tracked, a template matching technique, the sum of absolute difference (SAD), is used to locate the person in the collision area. The tracking is done after the person leaves the FOV of the camera. In the experimental studies, the tracking of one to three persons was performed using the implemented prototype. The average locating error of the tracking phase is only 5 pixels. The highest and average tracking success rates are 95.9that the proposed approach is accurate and feasible for people tracking by incorporating GPS and IP cameras.


Yang Koo Lee, Jungpil Shin, Kwang Deuk Kim, and Keun Ho Ryu. An Adaptive Data Reduction and Historical Query Processing for Storage-Centric Sensor Network. International Journal of Innovative Computing, Information and Control (SCI), 7(5):2945-2959, May 2011.

In this paper we propose an adaptive data reduction scheme and historical query processing technique based on a proxy architecture for the storage-centric sensor network towards efficient energy consumption. In our approach, sensor nodes maintain the collected data based on a novel three-phase summarization strategy. The major idea is that for a specific time period if all sensor reading values are not too much different from each other, appropriate representative values will be generated. As for query processing, we employ a proxy to transform user queries whose range criteria overlap into simpler fragment queries, each with a distinctive spatial and/or temporal predicate. In this way, only queries that have never appeared before are disseminated to sensor network. Consequently, the cost of communication and processing queries in sensor network is remarkably diminished. Furthermore, the experimental results validate that in spite of approximate sensor values obtained by the data summarization method responses to queries are still highly accurate.


T. Morota, J. Haruyama, M. Ohtake, T. Matsunaga, C. Honda, Y. Yokota, J. Kimura, Y. Ogawa, N. Hirata, H. Demura, A. Iwasaki, T. Sugihara, K. Saiki, R. Nakamura, S. Kobayashi, Y. Ishihara, H. Takeda, and H. Hiesinger. Timing and characteristics of the latest mare eruption on the Moon. Earth and Planetary Science Letters, 302:255-266, 2011.

Unraveling the timing and duration of mare volcanism on the Moon is essential for understanding its thermal evolution. The end of mare volcanism is poorly constrained, because mare basalts are incompletely sampled. In this study, employing SELENE (Kaguya) high-resolution images, we performed new crater size-frequency measurements for 49 young mare units (<3.0 Ga) in the Procellarum KREEP Terrane (PKT), in which the latest magma eruption of the Moon occurred. Mare volcanism in this region continued until 1.5 Ga, suggesting that volcanic activity in this region ceased 1.0 Ga after the magma eruption had globally ceased 2.5-3.0 Ga. Volcanic activity may have peaked 1.8-2.2 Ga ago. The youngest basalts occur around the Aristarchus plateau and the Kepler crater, which are located in the central region of the PKT. It is likely that heating in the crust due to the concentration of heat-producing elements in the PKT delayed cooling of a partial-melting zone in the underlying mantle. In contrast with previous basalt dating in this region, our results indicate a higher correlation between ages and spectral types of mare basalts; the young mare units in the PKT tend to have spectral types corresponding to high titanium contents, while low titanium basalts occur mainly in the early stage. The titanium variation in mare basalts may reflect vertical heterogeneity in TiO2 content in the upper mantle beneath the PKT.


Y. Ogawa, T. Matsunaga, R. Nakamura, K. Saiki, M. Ohtake, T. Hiroi, H. Takeda, T. Arai, Y. Yokota, S. Yamamoto, N. Hirata, T. Sugihara, S. Sasaki, J. Haruyama, T. Morota, C. Honda, H. Demura, K. Kitazato, J. Terazono, and N. Asada. The widespread occurrence of high-calcium pyroxene in bright-ray craters on the Moon and implications for lunar-crust composition. Geophysical Research Letters, 38(17):L17202L17202, 2011.

We investigated the continuous spectral features of fresh craters on the Moon accompanied by distinctive bright rays, with cavity diameters between 8 and 24 km. We used the data from the Spectral Profiler onboard SELENE (Kaguya) to gain a better understanding of the composition of the lunar highland crust. We found that the observed spectra exhibited strong symmetric absorption around 1 spectra exhibited strong symmetric absorption around 1 ?m and recognizable m and recognizable absorption around 1.3 absorption around 1.3 ?m. The spectra around a few craters showed a drastic change m. The spectra around a few craters showed a drastic change in the relative strengths of these two absorption bands s1:3=1:0 at different locations in and around the craters, indicating differences in the abundance of plagioclase and mafic minerals. In contrast, the spectra around most of the craters showed no significant variation in spectral shape, with an essentially constant s1:3=1:0. We analyzed the absorption features of the craters with an essentially constant s1:3=1:0 using the Modified Gaussian Model. We found that the strongest symmetric absorption bands were centered at 0.97-1.01 centered at 0.97{1.01 ?m with s1:3=1:0 ss 0.2{0.6. Comparing these values with data m with s1:3=1:0 ss 0.2-0.6. Comparing these values with data from known samples, we concluded that high-calcium pyroxene (HCP) is the most plausible dominant mafic mineral identified from the observed spectra. The fact that we detected such HCP-dominant spectra among rayed craters widely spaced across the lunar highland implies that the major mafic component of some portions of the lunar crust is HCP rather than low-calcium pyroxene (LCP).


S. Yamamoto, R. Nakamura, T. Matsunaga, Y. Ogawa, Y. Ishihara, T. Morota, N. Hirata, M. Ohtake, T. Hiroi, Y. Yokota, and J. Haruyama. Olivine-rich exposures in the South Pole-Aitken Basin. Icarus, 218:331-344, 2012.

The distribution and the geological context of the olivine-rich exposures in the South Pole-Aitken (SPA) Basin on the Moon were investigated based on the spectral data obtained from the Spectral Profiler (SP) and Multiband Imager (MI) onboard the Japanese lunar explorer Kaguya/SELENE. The olivine-rich exposures are found only in the peak rings or central peaks of the Schrdinger basin and Zeeman crater, which are located in the outer region of the SPA Basin and not in the center region. On a localized scale, the olivine-rich materials are exposed on landslide features on the crater walls or sloped wall of the central peaks or the peak rings. Another observational finding is the co-existence of olivine-rich and plagioclase-rich materials on a kilometer scale spanning most of the olivine-rich sites in the Schrdinger basin. Pyroxene-rich materials are found in fresh craters outside the peak rings or the central peaks with olivine-rich materials. Based on these results, the following scenario are proposed: (1) the impact to form the SPA Basin melted a large amount of the lunar upper mantle and crust, and distributed the melted materials to the outer region; (2) local differentiation of melted materials hid the olivine-rich materials in the center region of the SPA Basin; (3) later impacts that formed the Schrdinger and Zeeman craters excavated and exposed the olivine-rich materials to the surface again; and (4) space weathering and regolith gardening obscured the olivine-rich spectra at the exposure sites, but recent, small scale impacts or landslides on the sloped wall exposed fresh olivine-rich materials, allowing the identification of the olivine-rich exposures by spectral remote-sensing. This suggests that several, different scale events play an important role in forming the surface distributions of originally deep-seated materials on the Moon, as well as on other planetary bodies.


Y. Yokota, T. Matsunaga, M. Ohtake, J. Haruyama, R. Nakamura, S. Yamamoto, Y. Ogawa, T. Morota, C. Honda, K. Saiki, K. Nagasawa, K. Kitazato, S. Sasaki, A. Iwasaki, H. Demura, N. Hirata, T. Hiroi, R. Honda, Y. Iijima, and H. Mizutani. Lunar photometric properties at wavelengths 0.5-1.6 um acquired by SELENE Spectral Profiler and their dependency on local albedo and latitudinal zones. Icarus, 215(2):639-660, 2011.

The lunar photometric function, which describes the dependency of the observed radiance on the observation geometry, is used for photometric correction of lunar visible/near-infrared data. A precise photometric correction parameter set is crucial for many applications including mineral identification and reflectance map mosaics. We present, for the first time, spectrally continuous photometric correction parameters for both sides of the Moon for wavelengths in the range 0.5-1.6 um and solar phase angles between 5 deg and 85 deg, derived from Kaguya (SELENE) Spectral Profiler (SP) data. Since the measured radiance also depends on the surface albedo, we developed a statistical method for selecting areas with relatively uniform albedos from a nearly 7000-orbit SP data set. Using the selected data set, we obtained empirical photometric correction parameter sets for three albedo groups (high, medium, and low). We did this because the photometric function depends on the albedo, especially at phase angles below about 20 ° for which the shadow hiding opposition effect is appreciable. We determined the parameters in 160 bands and discovered a small variation in the opposition effect due to the albedo variation of mafic mineral absorption. The consistency of the photometric correction was checked by comparing observations made at different times of the same area on the lunar surface. Variations in the spectra obtained were lower than 2the large phase angle data in mare. Lastly, we developed a correction method for low solar elevation data, which is required for high latitude regions. By investigating low solar elevation data, we introduced an additional correction method. We used the new photometric correction to generate a 1 deg mesh global lunar reflectance map cube in a wavelength range of 0.5-1.6 um. Surprisingly, these maps reveal that high latitude (>75 deg) regions in both the north and south have much lower spectral continuum slopes (color ratio r1547.7nm/r752.8nm<1.8) than the low and medium latitude regions, which implies lower degrees of space weathering.

Refereed Proceedings Papers


Dang Tuan Anh, N. Hirata, N. Asada, and H. Demura. Method for Texture Mapping on Irregular Shaped-Bodies in GIS Oriented Applications. In 11th Asian Symposium on Visualization (ASV11), page 218, Niigata, June 2011.

The Visualization Society of Japan. 3D-GIS Web Application for irregular-shaped asteroids is a tool for planetary scientists to study asteroids. Users can analyze scientific data on the asteroid surface mapped to a polygon model (3D triangle mesh). Because current mapping data are as attributes of each triangle on the model, their resolutions depend on that of the model. There is no efficient way to map data with a different resolution from that of the mesh. Here we propose a novel method to solve the problem. Because asteroids have the same topology of a sphere, it would be natural if we apply spherical parameterization, i.e., assigning a 3D position on the unit sphere to each of the mesh vertices. Then, we apply traditional cylindrical projection to make the texture map. Such maps will be independent on models.


S. Sato, N. Asada, H. Demura, and N. Hirata. Extraction of craters as concentric circle patterns of contours on the lunar digital terrain model. In IEEE International Geoscience and Remote Sensing Symposium (IGARSS 2011), page 3803, Vancouver, Canada, July 2011. IEEE.

Impact crater is one of the important topographic features on the planet. Sizefrequency distributions of craters on the plan etary surfaces are used for determining planetary chronology. This paper presents an algorithm for automatic crater detec tion to reduce time to counting craters. A crater is a circu lar depression. So contour lines of a crater on a topographic model form a characteristic pattern of concentric circles. Our algorithm detects craters in digital terrain model (DTM) with this concept. There is no need to care the angle of the solar incident because the DTM is independent on the solar alti tude. An evaluation test with a lunar DTM shows that 80 are extracted. So there is possibility to extract most of craters, including small size of craters such as them, on the planet by more improvement of the algorithm. Most false detections are caused by unintended detections of circular positive reliefs, non-crater circular de pressions and other complex topographic features.


Taewook Kim, Meijing Li, Keun Ho Ryu, and Jungpil Shin. Prediction of Protein Function from Protein-Protein Interaction Network by Weighted Graph Mining. In International Conference on Bioinformatics and Biomedical Technology (ICBBT 2012), pages 26-28, Singapore, Feb. 2012. CBEES, CBEES.

Protein-protein interaction network plays a key role in protein function prediction. Many previous studies attempted to assign the protein function by neighbor or the connected path way to known protein. However, since interacting proteins do not always correspond, their accuracy is limited. In this paper, we proposed a novel approach to predict the protein function by weighted graph mining. Our proposed approach finds the functional frequent 2-node and 3-node patterns by labeling each protein node as a set of corresponding functions to predict the protein function without functional inconsistent between interacting protein. It also makes a selection from discovered frequent patterns by applying the weight to each edge so that we could compare which pattern is the most reliable for the function prediction. The function prediction is performed by matching the selected 2-node patterns, interacting with unannotated protein, with the frequent 3-node patterns. In the experiment, we used yeast protein interaction network which has functional frequent 3-node patterns and result shows 0.653 of function prediction accuracy. Compared with other methods’performance, our approach is better.


Yu Tang and Jungpil Shin. De-ghosting for Image Stitching with Automatic. In Aytul Ercil, editor, IEEE 20th International Conference on Pattern Recognition (ICPR2010), pages 2210-2213, Istanbul, Turkey, Aug 2010. IEEE, IEEE.

Ghosting artifact in the field of image stitching is a common problem and the elimination of it is not an easy task. In this paper, we propose an intuitive technique according to a stitching line based on a novel energy map which is essentially a combination of gradient map which indicates the presence of structures and prominence map which determines the attractiveness of a region. We consider a region is of significance only if it is both structural and attractive. Using this improved energy map, the stitching line can easily skirt around the moving objects or salient parts based on the philosophy that human eyes mostly notice only the salient features of an image. We compare results of our method to those of 4 state-of-the-art image stitching methods and it turns out that our method outperforms the 4 methods in removing ghosting artifacts..


Donggyu Lee and Jungpil Shin. Discovery of Association Rules with High Confidence from Korea Acute Myocardial Infarction Registry. In The 4th International Conference on the Frontiers of Information Technology Application and Tools (FITAT2011), pages 94-96, Chugnbuk, Korea, May 2011. KIPS, KIPS.

Korean Acute Myocardial Infarction Registry (KAMIR) is on-line database of patients with Acute Myocardial Infarction(AMI) supported from the Korean Society of Cardiology since November 2005. In this paper, we mine meaningful rules from KAMIR database using the Apriori algorithm. Our goal is to find causality between high risk of cholesterol, hs-CRP, and smoking, and patients with hypertension, diabetes mellitus, and hyperlipidemia.


Takashi Watanabe, Shingo Watanabe, and Jungpil Shin. Writer Identification using the Wii Remote Controller. In IEEE 3rd International Conference on Awareness Science and Technology (IEEE iCAST 2011), pages 27-30, Dalian, China, Sep 2011. IEEE, IEEE.

The objective of this study was to develop a system for handwriting recognition in three dimensions (3D) to authenticate users. While previous studies have used a stylus for twodimensional input on a tablet, this study uses the Wii Remote controller because it can capture 3D human motion and could therefore be more effective. The information obtained from a Wii Remote controller included x and y coordinates, acceleration (x, y, z), angular velocity (pitch, yaw, roll), twelve input buttons, and time. The proposed system calculates distances using six features extracted after preprocessing the data. In an experiment where 15 subjects wrote “ AIZU ”10 times, we obtained a 94.8rate using a combination of writing velocity, the peak value of pitch, and the peak value of yaw. This suggests that this system holds promise for handwriting-based authentication in the future.


Junichi Sato and Jungpil Shin. Signature Verification for Multiuser Online Kanji Learning System. In Jung-Young Son, editor, 5th 2011 International Universal Communication Symposium (IUCS2011), pages 12-14, Gumi, South Korea, Oct 2011. KIPS, KIPS.

In this paper, we propose to reject the possibility to accept the login by someone else, which is not by himself or herself at multiuser online kanji learning system with signature verification. Human’s signature is one of the human's biometrics. Biometrics includes very personal information and characteristics of usefulness, because it is human’s looking or behavior feature. By signature verification, we could verify the writer is proper or not with human’s writing behavior, so dishonest writer will be rejected. Using signature verification as self verification at the system, we could consider two advantages. One advantage is advancement of security which will cause from deterrent of someone else’s scamming. The other advantage is advancement of usefulness which will cause from decrease of the number of using input device, only pen device will be need to handle the system. We adopt signature verification as calculating similarity by using some reference signature and the distance which will calculated by DP matching in this research. Input signature’s self similarity will be calculated by dividing the average distance between input and each reference signature data with average distance between each reference signature data. From signature verification’s experimental results which changes using features, we adopted to use writing velocity and writing speed differential as using feature to verify the writer for the system. By using signature database which is construct with 20 genuine signatures and 20 forged signatures with 40 writers and written mostly by English or Chinese literal, experimental results of signature verification records 12.71maximum EER, 6.00when we establish the threshold as constant, to simulate the signature verification function is implement in actual running system, average of FRR records 10.29advance the reliability and usefulness of the multiuser online kanji learning system.


Seo In Jeon, Taewook Kim, Meijing Li, Ho Sun Shon, Keun Ho Ryu, and Jungpil Shin. Frequent Pattern Mining for Protein Function Prediction in Protein-Protein Interaction Network. In the 2012 International Conference on Convergence Technology (ICCT 2012), pages 35-37, Hochiminh, Vietnam, Jan 2012. KCS, KCS.

Predicting protein function from protein-protein interaction network is the most challenging problem. Because of its complexity and diversity, it is still remaining difficult task. In this paper, we proposed novel method to define protein functions from protein-protein interaction network. We represent the protein interaction network as a graph and using graph mining approach to detect frequent patterns. We find frequent 2-node patterns and predict the protein functions using them. To evaluate our method, we compare our method with neighbor counting method using same data set. The data set which we used is obtained from DIP. The result shows that the performance of our method is better than Neighbor counting method.


Tsendsuren Munkhdalai, Meijing Li, Taewook Kim, Oyun-Erdene Namsrai, Seon phil Jeong, Jungpil Shin, and Keun Ho Ryu. Bio Named Entity Recognition based on Co-training Algorithm. In Fifth International Symposium on Mining and Web, Fukuoka, Japan, Mar 2012. KIPS, KIPS.

One essential task in extracting information from biomedical literature is the bio Named Entity Recognition (NER) process, which basically defines the boundaries between typical words and biomedical terminology in particular text data, and assigns them based on domain knowledge. This paper presents a semisupervised integration of completely different classifiers to cover knowledge from unlabeled data to recognize bio named entities in text. We modified the original co-training, a semisupervised learning algorithm, with a scalable feature processing schema, which extracts the bio NER feature from a number of unlabeled data and converts different types of feature sets. Our base result shows that the classifiers of co-training achieve significant learning from unlabeled data.



Jungpil Shin. Laser based Intelligent Input Device for Interactive Screen, June 2011.

Co-project with Seoul University in Korea and the Univ. of AIZU

Academic Activities


N. Asada, June 2011.

Scientific Adviser, ASV11, the Visualization Society of Japan


Jungpil Shin, Nov 2011.

Designated Reviewer (held in Kuala Lumpur, Malaysia, Nov. 16-18 2011)


Jungpil Shin, Jan 2012.

Program Committee, (held in, Ho Chi Minh, Vietnam, January 26-28, 2012)


Jungpil Shin, Apr 2011.

Member of the Program Committee (held in Orlando, Florida US, April, 2011), the 2011 International Conference on Knowledge Generation, Communication and Management (KGCM 2011)


Jungpil Shin, May 2011.

Chair of Workshop (held in Cheongju Chungbuk National University, Korea, May 1113, 2011), The 4th International Conference on Frontiers of Information Technology, Application and Tools (FITAT 2011) and The 2nd International Workshop on Data Mining Applications: BioInformatics and Medical Information (DBMI 2011)


Jungpil Shin, Dec 2011.

Program Committee,The 2nd International Conference on Engineering and Industries (ICEI2011) (held in Jeju, Korea, Nov 29 - Dec 1, 2011)


Jungpil Shin, Sep 2011.

Publication Committee Chair, IEEE 3rd International Conference on Awareness Science and Technology (IEEE iCAST 2011),(held in Dalian, China, Sep. 27-30, 2011)


N. Hirata, 2011.

Member, and member of the committee for general affairs, the Japanese Society for Planetary Sciences


N. Hirata, 2011.

Member of Local Organizing Committee, Asteroid, Comet, Meteor meeting 2012


N. Hirata, 2011.

Member of Program Subcommittee, the 28th International Symposium on Space Technology and Science


N. Hirata, 2011.

Member of Program Subcommittee, the 29th International Symposium on Space Technology and Science

Ph.D., Master and Graduation Theses


Tomoetsu Sugawara. Web-GIS Application for Analysis of Reflectance spectra of Surface Materials on the Moon. Master thesis, Graduate School of Computer Science and Engineering, 2012.

Thesis Adviser: N. Hirata


Dang tuan Anh. Method for Texture Mapping on Irregular-Shaped Bodies in GIS Oriented Applications. Master thesis, Graduate School of Computer Science and Engineering, 2011.

Thesis Adviser: N. Asada


Kotomi Ishida. Verification of Educational Effect of Kanji Learning System. Graduation Thesis, School of Computer Science and Engineering, 2011.

Thesis Adviser: Jungpil Shin


Kazunari Soma. On-Line Signature Evaluation Using Fuzzy Rules. Graduation Thesis, School of Computer Science and Engineering, 2011.

Thesis Adviser: Jungpil Shin


Makoto Marumoto. Realization of Ink Diffusion Phenomenon in 3D Virtual Calligraphy Simulatio. Graduation Thesis, School of Computer Science and Engineering, 2011.

Thesis Adviser: Jungpil Shin