Papers - THI THI ZIN
-
A stochastic model for web reliability ranking system Reviewed
Pyke Tin, Thi Thi Zin, T. Toriu, H. Hama
ICIC Express Letters 4 ( 3 ) 705 - 711 2010.6
Language：English Publishing type：Research paper (scientific journal) Publisher：ICIC Express Letters
The World Wide Web is used, by an increasing number of people as an ever expanding source of information on almost every topic imaginable. However, useful data, is often buried, in large quantities of low-reliability content. Estimation of Web information reliability is valuable for diverse applications, such as a, search result ranking and, a, direction of crawlers. In this paper, we propose a, novel stochastic model for web information ranking system, which enables us to search useful and, reliable knowledge information. Specifically, a multiple state reliability ranking model based on the theory of Markov chain is developed, by assuming that the likelihood, of a, statement on the Web can be trusted, using standards developed, by information scientists, and, the link structure of associated, web pages. We then cluster relevant and, reliable webpages based, on whether they can be trusted, or not. Finally, the proposed model is tested, on an academic search engine and, show how the reliability ranks can be used, for searching a, useful knowledge.
-
Human behaviors analysis at or near public transportation asset
Thi Thi Zin, K. Fujimura, S. Kamijo
17th ITS World Congress 2010
Authorship：Lead author Language：English Publishing type：Research paper (scientific journal) Publisher：17th ITS World Congress
Security of human lives and property has always been a major concern for civilization for several centuries. In modern civilization, the threats of theft, accidents, terrorists' attacks and riots are ever increasing in public access areas such as airports, train stations, shopping malls, banks and etc. Due to the high amount of useful information that can be extracted from a video sequence, video surveillance has come up as an effective tool to forestall these security problems. In surveillance systems, understanding human behaviors and activities arising out of the interactions of various objects, as well as their evolution over time is an important problem. The protection of critical transportation assets and infrastructure is an important topic in these days. In this paper, we develop a new rule based approach to smart video surveillance system for detecting situations where people may be in peril, as well as suspicious action or interactions at or near critical transportation assets. For organizational purposes, the surveillance operationally-relevant behaviors are divided into three general groups: (i) single person or no interaction, (ii) multiple person interactions, and (iii) person-facility/location interactions. The behavior analysis is accomplished through the development of geometric and motion visual features for each pedestrian. With this information, the system could alert authorities if pedestrians display suspicious behaviors. The performance evaluation of the proposed system is carried out by using the video sequences taken in the real life environments of rail stations. The experimental results show the high accuracy rates.
-
Optimal crawling strategies for multimedia search engines Reviewed
H. Hama, Thi Thi Zin, Pyke Tin
IIH-MSP 2009 - 2009 5th International Conference on Intelligent Information Hiding and Multimedia Signal Processing 182 - 185 2009.12
Language：English Publishing type：Research paper (international conference proceedings) Publisher：IIH-MSP 2009 - 2009 5th International Conference on Intelligent Information Hiding and Multimedia Signal Processing
In this paper we propose a novel optimal crawling strategy for next-generation multimedia search engines. We consider here a Web crawl as a two-dimensional (2D) random walker on a graph whose vertices are the Web pages and whose edges are the hyperlinks. The proposed crawler is a two-part scheme optimizing the crawling process in such a way that the average level of staleness over all pages is minimized and the quality of search engine from user's perspective is maximized. In doing so, we employ techniques from probability theory and the theory of functional equations which are highly computationally efficient-crucial for practicality because the size of the problem in the Web environment is immense. We show that a combination of breadth-depth crawling including the largest sites is a practical and optimal strategy. In particular, several probabilistic models for user browsing in infinite Web are proposed and studied to estimate how deep and breadth a crawler must go to download a significant portion of the Web site that is actually visited. Experimental and simulation results show that a crawler needs to download just a few levels in depth and breadth to reach the maximum number of pages that users actually visit. It also suggests that the largest sites should be included in the crawling process. © 2009 IEEE.
-
Dominant color embedded Markov chain model for object image retrieval
Zin T., Tin P., Toriu T., Hama H.
IIH-MSP 2009 - 2009 5th International Conference on Intelligent Information Hiding and Multimedia Signal Processing 186 - 189 2009.12
Language：Japanese Publishing type：Research paper (scientific journal) Publisher：IIH-MSP 2009 - 2009 5th International Conference on Intelligent Information Hiding and Multimedia Signal Processing
This paper proposes a new and compact method for object image retrieval fusing Dominant Colors (DCs) and embedded Markov chain concepts. This proposed method uses combined color-texture features which are characterized in terms of their spatial interaction or interrelationship properties, modeled by means of a set of embedded Markov chains, each associated with a major spatial direction. Specifically, DCs are extracted from the object image, which are encountered pixel-wise along a given direction to form an embedded Markov chain. Normalizing the resultant Markov chains over all specified directions, the corresponding stationary distribution is derived and served as Markov Feature-Vector (MFV). We then employ the chi square distance between the feature vectors in comparing similarity of images. The MFV involves spatial structure information of both within and between dominant color regions. Moreover, it keeps simplicity, compactness, efficiency, and robustness. We conduct experiments using a comprehensive set of images including deformable shapes. Experimental results show that the proposed method can retrieve an important number of correct images with very high accuracy while the mismatch ratio remains constant. © 2009 IEEE.
-
Bundling multislit-HOG features of near infrared images for pedestrian detection
Zin T., Tin P., Hama H.
2009 4th International Conference on Innovative Computing, Information and Control, ICICIC 2009 302 - 305 2009.12
Language：Japanese Publishing type：Research paper (scientific journal) Publisher：2009 4th International Conference on Innovative Computing, Information and Control, ICICIC 2009
In this paper we present a novel scheme where image features are bundled into local groups. Specifically, features of Near Infrared (NIR) images extracted by using Histogram of Oriented Gradients (HOG) descriptor and those by our multislit method are bundled into a single descriptor. The method involves first localizing the spatial layout of body parts (head, torso, and legs) in individual frames using multislit structures, and associating these through a series of extracting HOG features. A bundled feature vector describing various types of poses is then constructed and used for detecting the pedestrians. Experiments with a database of NIR images show that our scheme achieves a substantial improvement in average precision over the baseline conventional HOG approach. Detection and recognition performance is less computationally expensive than existing approaches. © 2009 IEEE.
-
An optimal choice of morphological operating center for object image retrieval
Hama H., Zin T., Tin P.
2009 4th International Conference on Innovative Computing, Information and Control, ICICIC 2009 298 - 301 2009.12
Language：Japanese Publishing type：Research paper (scientific journal) Publisher：2009 4th International Conference on Innovative Computing, Information and Control, ICICIC 2009
In this paper we introduce a novel and simple schemes to develop an optimal choice of morphological Operating Center (OC) for object image retrieval. A variation of the standard morphological operators which require the choice of an OC is discussed. The proposed method is based on combinations of statistical and dynamic programming techniques in which recursive equations on the basis of dilation by using the principle of optimality and minimizing unnecessary background area of the objects are applied. We also present the application of mathematical morphology with Structuring Elements (SEs) which are elongated in the angular direction. The experimental results show that the optimal choice of OC provides satisfying retrieval results. © 2009 IEEE.
-
Reliability based web information ranking system
Tin P., Zin T., Toriu T., Hama H.
2009 4th International Conference on Innovative Computing, Information and Control, ICICIC 2009 294 - 297 2009.12
Language：Japanese Publishing type：Research paper (scientific journal) Publisher：2009 4th International Conference on Innovative Computing, Information and Control, ICICIC 2009
In this paper, we propose a reliability based Web information ranking system which enables searching useful and reliable knowledge information. The proposed system will contain subsystems for reliability ranking, information clustering based on reliability. The reliability ranking system will estimate the likelihood that a statement on the Web can be trusted using standards developed by information scientists, and the link structure of associated Web pages. The clustering will cluster relevant and reliable information based on whether or not they can be trusted or not. We test these models on an academic search engine and show how the reliability information ranks can be used as a useful knowledge. © 2009 IEEE.
-
Spatial image retrieval based on dynamic thresholding
Zin T., Hama H., Tin P.
International Journal of Innovative Computing, Information and Control 5 ( 11 ) 4051 - 4059 2009.11
Language：Japanese Publishing type：Research paper (scientific journal) Publisher：International Journal of Innovative Computing, Information and Control
In this paper, we present a spatial image retrieval method based on dynamic thresholding. The proposed method can retrieve spatial image patterns with high accuracy and speed from images with complicated backgrounds. For simplicity, we consider the query images as specified rectangular-shaped or circular-shaped framed images. First, by introducing a dynamic thresholding system, the images can be partitioned into Peak Color Regions (PCRs). Consequently, the proposed method requires low computational complexity giving optimal feasible results for detection and segmentation. Due to compact representation and low complexity of color features, direct histogram comparison is to be used for extraction of PCRs. Since the number of the PCRs is much smaller than that of the image pixels, the proposed method allows a low dimensional image processing. The effectiveness of the proposed method is confirmed through experiments with various images. © 2009 ISSN.
-
A hybrid ranking of link and popularity for novel search engine Reviewed
H. Hama, Thi Thi Zin, Pyke Tin
International Journal of Innovative Computing, Information and Control 5 ( 11 ) 4041 - 4049 2009.11
Language：English Publishing type：Research paper (scientific journal) Publisher：International Journal of Innovative Computing, Information and Control
In this paper, we explore a new paradigm to enable Web search at image level reflecting the most relevant results to the users. So we introduce a new concept of Popularity-based Rank (PR), a content-level ranking and searching model for image retrieval. Specifically, we establish a PR Operation which is combined with new link structure analysis. The strategy to decide searching order by taking similarity into consideration is also proposed and proved to be effective and efficient. Experimental results show that the combined analysis can achieve significantly better ranking results than naively applying page-level ranking on the image model which is usually used in the current search engines. © 2009 ISSN.
-
Zin T., Takahashi H., Hama H.
International Journal of Innovative Computing, Information and Control 5 ( 3 ) 751 - 761 2009.3
Language：Japanese Publishing type：Research paper (scientific journal) Publisher：International Journal of Innovative Computing, Information and Control
Nowadays, person detection in far infrared (FIR) images toward realizing a night vision system becomes a hot topic. However, sufficient performance could not be achieved by conventional schemes. Since the properties of FIR images different from visible images, it is not known what kind of scheme is appropriate for person detection in FIR images. In this paper, we propose two novel methods for person detection using FIR images: (i) body parts detection method and (ii) Gravity Center (GC) movement pattern method. First, we introduce the multi-slit method along with vanishing line for extraction of head regions. After the head region is detected and segmented, the person body and legs regions are roughly estimated by size ratios. The histograms of Sobel edge of such estimated regions are used to confirm the segmented head. This method can be applicable to person detection at both near and far distances in indoor and outdoor scenes. Second, we propose a sequential decision method by investigating GC movement patterns. It is very simple and especially valid for images at near distances. Our experiments demonstrate the effectiveness of the proposed methods and the advantages in dealing with person detection. Finally, comparative study and further extendable potential applications of the proposed methods are pointed out to be focused in our future research. © 2009 ISSN.
-
Ranking system for image database using special type of markov chain
Hama H.
Proceedings - Digital Image Computing: Techniques and Applications, DICTA 2008 556 - 563 2008.12
Language：English Publishing type：Research paper (scientific journal) Publisher：Proceedings - Digital Image Computing: Techniques and Applications, DICTA 2008
In this paper, we explore and examine a new novel approach to image ranking systems based on some special types of Markov chain along with new concepts of popularity and relevancy measures for image database. To be specific, this approach introduces a family of special Markov chain models in which serial correlations are explicitly involved so that we can use them as correlations among the images. By using these models, we develop a ranking function for the image database. On the other hand, 'popularity' and 'relevancy' concepts are introduced and used for developing an alternative ranking function for the database. In the process of developing ranking functions we use a method of queue-based stochastic difference equations. We then blend two ranking functions, to propose a new ranking system for searching order of image database. Since the proposed ranking system considers concepts of correlations, popularity and relevancy altogether, it is beneficial to a modern search engine for investigating behavior and effects of those parameters on the search results. Some illustrative examples and simulation results are presented with reference to a real world application domain. © 2008 IEEE.
-
Zin T., Hama H., Tin P.
Proceedings - Digital Image Computing: Techniques and Applications, DICTA 2008 548 - 555 2008.12
Language：Japanese Publishing type：Research paper (scientific journal) Publisher：Proceedings - Digital Image Computing: Techniques and Applications, DICTA 2008
This paper proposes a new method for object retrieval in image and video databases. The proposed system uses the histogram based approach along with Angular Radial Representation (ARR). In addition, concepts of Dominant Colors (DCs) and morphological dilation using ring-shaped and fanshaped Structuring Elements (SEs) are also applied. It is found that the approach is not only invariant to rotation, translation and scaling but also valid for low resolution images and partial occlusion. The retrieval effectiveness of the proposed system is shown through experiments using a comprehensive set of images including deformable shapes. © 2008 IEEE.
-
Novel search engine: Combination of link and popularity rank for multimedia retrieval
Hama H., Zin T., Tin P.
3rd International Conference on Innovative Computing Information and Control, ICICIC'08 2008.9
Language：Japanese Publishing type：Research paper (scientific journal) Publisher：3rd International Conference on Innovative Computing Information and Control, ICICIC'08
In contrast with the current search engines that essentially do page-level ranking and searching, we are exploring a new paradigm to enable Web search at image level by introducing a new concept of Popularity-based Rank (PR). This paper introduces PR, a content-level ranking and searching model for multimedia retrieval. Specifically, we establish a PR Operation which is combined with new link structure analysis. The strategy to decide searching order by taking similarity into consideration is also proposed and proved to be effective and efficient. Experimental results show that the combined analysis can achieve significantly better ranking results than naively applying PageRank on the image model. © 2008 IEEE.
-
Robust detection and segmentation of images with tolerance
Zin T., Hama H., Tin P.
3rd International Conference on Innovative Computing Information and Control, ICICIC'08 2008.9
Language：Japanese Publishing type：Research paper (scientific journal) Publisher：3rd International Conference on Innovative Computing Information and Control, ICICIC'08
In this paper, we introduce a novel approach to retrieve significant spatial images from a collection of images. The query images of interest in this paper are specified rectangular-shaped and circular-shaped as framed photographs. By incorporating the advantages of adaptive thresholding and tolerant partitioning into peak color regions, the proposed method requires low computational complexity and it is very feasible for image detection and segmentation. Peak image regions are then extracted by using color histograms. Since the number of the peak regions is much smaller than that of the image pixels, this proposed method allows a low dimensional image processing. The effectiveness of the proposed method is confirmed through experiments with various images. © 2008 IEEE.
-
Robust person detection using far infrared camera for image fusion Reviewed
Thi Thi Zin, H. Takahashi, H. Hama
Second International Conference on Innovative Computing, Information and Control, ICICIC 2007 2008.2
Language：English Publishing type：Research paper (international conference proceedings) Publisher：Second International Conference on Innovative Computing, Information and Control, ICICIC 2007
In this paper we present a robust method for person detection using far infrared images. To extract initial nominated head regions, thresholding and morphological operations are applied using intensity information. Among these regions, some of wrongly extracted regions are removed using the pattern of person head based on the local maximums of Sobel edge image. After the head regions are segmented, the person body and legs region are roughly estimated by the ratios. The histograms of Sobel edge of such estimated regions are used to confirm the segmented head. This method can be applicable to person detection at both near and far distances in indoor and outdoor scenes. Moreover, we propose another novel algorithm using the movement pattern of gravity centers. It is a very simple way, especially valid for images at near distances. Our experiments demonstrate the effectiveness of the proposed method and the advantages in dealing with person detection for night vision applications. Finally, image fusion of visible and far infrared cameras is discussed.
-
Robust signboard recogniton in the presence of occlusion and reflection Reviewed
Thi Thi Zin, H. Hama, S-S. Koh
International Journal of Innovative Computing, Information and Control 3 ( 6 A ) 1321 - 1334 2007.12
Language：English Publishing type：Research paper (scientific journal) Publisher：International Journal of Innovative Computing, Information and Control
Recognizing objects for visual information in outdoor scenes is very useful but challenging. This paper presents a new framework to recognize signboards with uniform color regions in the presence of occlusion and reflection. The framework is composed of three stages: (i) extraction of uniform color regions by adaptive rank filter and piecewise linear approximation, (ii) recognition by template matching, and (iii) verification by relative color polygons. In the experiments, we used 300 images taken under a great variety of adverse conditions including occlusion, reflection, specular highlights, and so on. The proposed system achieved 98% recognition rate for images taken under such bad conditions. Moreover, it can be extended for vision-based car and pedestrian navigation systems to provide up-to-date information to the users and potentially be embedded in a driver assistance system.
-
Accurate reconstruction of non-rigid 3D shapes Reviewed
S-S. Koh, Thi Thi Zin, H. Hama
Digest of Technical Papers - IEEE International Conference on Consumer Electronics 2007.8
Language：English Publishing type：Research paper (international conference proceedings) Publisher：Digest of Technical Papers - IEEE International Conference on Consumer Electronics
The aim of this paper is to create an accurate 3D-TV contents from 2D-video which can be used in a commercial 3D-TV broadcast environment. In this paper we propose an accurate recovering of non-rigid 3D shapes using a chained partial rigid factorization method. We have extended the existing rigid factorization algorithm to the stereo camera case. Our focus in this paper is on estimating deformable shapes from stereo images. Our experiments show the improved results for reconstructions. ©2007 IEEE.
-
Robust signboard recognition for vision-based navigation Reviewed
Thi Thi Zin, S-S. Koh, H. Hama
Kyokai Joho Imeji Zasshi/Journal of the Institute of Image Information and Television Engineers 61 ( 8 ) 1192 - 1200 2007.8
Language：English Publishing type：Research paper (scientific journal) Publisher：Kyokai Joho Imeji Zasshi/Journal of the Institute of Image Information and Television Engineers
We propose a framework that enables vision-based navigation systems to more robustly recognize signboards. It is designed mainly for signboards with uniform color regions, which can be strongly affected by occlusion and reflection. The framework comprises two stages: preprocessing for extracting uniform color regions and recognizing them. In the first stage, the cumulative frequency, morphology gradient, and Mahalanobis distance are used for the extraction. In the second stage, a new matching method for identifying templates and a new relative method for verifying colors are used. Testing of a system using this framework and 300 images taken under various conditions showed a recognition rate of 98%, demonstrating the framework's powerful ability to recognize signboards.
DOI： 10.3169/itej.61.1192
-
Optimal color space for relative color polygons
Zin T., Koh S., Hama H.
IEICE Electronics Express 4 ( 3 ) 106 - 113 2007.3
Language：Japanese Publishing type：Research paper (scientific journal) Publisher：IEICE Electronics Express
We aim for detection and recognition of planar objects in natural outdoor scenes under varying illumination conditions. To achieve this, we propose Relative Color Polygons (RCPs) using component colors of objects for color matching. They can be defined on many color spaces, and it is found that a 2D color space (XY space) is the optimal color space for our relative color method compared with other color spaces. To evaluate the invariance to illumination changes for object recognition, experiments have been carried out using 500 outdoor scene images. By using the proposed model, the color matching rate of the input images with the standard one was 95%. This framework is potentially applicable to image retrieval, image segmentation, image recognition, and so on. © IEICE 2007.
DOI： 10.1587/elex.4.106
-
Geometrical properties between 2-D image plane and 3-D error space
Koh S., Zin T., Hama H.
IEICE Electronics Express 3 ( 14 ) 333 - 339 2006.7
Language：Japanese Publishing type：Research paper (scientific journal) Publisher：IEICE Electronics Express
We study geometrical properties between 2-D image plane and 3-D error space under affine reconstruction. The purpose of our system is to contribute to more accurate 3-D reconstruction by analyzing geometrically 2-D to 3-D relationship. In situation for no missing feature points and no noise in the 2-D observation matrix, the accurate solution is known to be provided by Singular Value Decomposition. However, several feature points of the matrix have not been observed because of occlusions and low image resolution, and so on. In this case, there is no simple solution. To obtain accurate 3-D reconstruction by recovering missing feature points, we propose the analytic approach which can handle the error orientation and distance of missing feature points by the geometrical properties. © IEICE 2006.
DOI： 10.1587/elex.3.333