[The state of the art in multi-media search]
●Web Video
- Google Video, Youtube : post type
-- Tagging by people who post the materials
-- Search by tag words
- Gyao, Yahoo video: WebTV type
-- Tagging by RSS etc to each video data
-- Manual categorization, time-limited Upload
●Image
- Google Image, MSN Image Search
-- Text/word matching in the attached text or caption
-- Using the results of semantic analysis
- Mixi, Blog etc
-- Manual tagging
●Sound, Music
-- Manual tagging, attachment of meta-information by media sellers
-- Search by MIDI code data, melody search using note sequence
-- Retrieval using phoneme sequence and CDP (The upper parliament video publication (an application by our method)
[Multimedia Search Engine]
●System realized based on Flex + C + MySQL（or Ajax + PHP + C + DB）
- Engine is described by C-native
- Storing ontology data in DB
- Broadcasting XML information using REST
→available to re-use
- Proposal a new frontier of multi-media search in the Web
- Knowledge integration based on textual and no-textual information in the WEB2.0 paradigm
- An organized project in the Univ of Aizu opened to the public
Non indexed speech and image signal data in Youtube-like Web is accessed by signal and/or text query
- Speech and song retrieval(phoneme labelling and CDP spotting)
- Image and video retrieval（partial and similar image detection,video summary and character recognition）
- Ontology construction using query tags
- Providing our RSS
Legacy information systems are typically the backbone of organizations and their failure can be crucial to their businesses. Legacy information systems can cause various problems since, for example, they usually run on obsolete hardware that is slow and expensive to manage, the software is developed using obsolete technologies that make integration with new systems and tracking of faults difficult, documentation and resources for maintenance are lacking. Several solutions have been proposed to these problems, such as, redevelopment, wrapping, and migration. However, given a concrete legacy information system, it is not always possible to rely solely on one solution.
In every legacy information system problem, first an intensive study to find the most appropriate approach must be performed. Since this area of software engineering is relatively new, the numbers of practical experience reports of a comprehensive migration approaches are limited and sound methodologies do not exist.
The aim of this project is to devise a comprehensive methodology for renovating a legacy system. An actual, mission critical legacy information system currently used in a relatively large organization will be considered. An intensive study to find the most appropriate approach for solving the organization's problem will be performed as well as research of existing knowledge in literature. A proposal of the methodology backed by analysis of risks and impact on business will be the output of this project.
1 Analysis & study (2 months) Report of literature, Target description
2 Requirements spec. (1month) Requirements specification
3 Basic design (1 month) Basic design report
4 Detailed design (3 months) Detailed design report
5 Simulation (2 months) Software, cost study report
6 Proposal prep. (1 months) Renovation proposal
7 Methodology (2 months) Thesis report
- Necessary software: will be decided within project.
- Necessary hardware: will be decided within project.
- Software Engineering:
Prior job experience will be great. For students, an undergraduate course on software engineering will be fine (Not a prerequisite but a strong recommendation).
- Databases:
Corresponding knowledge and skills are recommended. If necessary, students will be requested to take corresponding courses or study literature.
This project is related to the following undergraduate courses.
(These are NOT pre-requisites.)
a. Software Engineering I
b. Software Engineering II
c. Web Programming
d. Database
Students are recommended to take the following courses from the educational curriculum
a. Requirements Engineering
b. Software project management
ソフトウエア開発における品質の問題は、現在、ソフトウエア開発に携
わる全ての会社で発生しているといって過言でなく、経済紙等で言うと
ころでの社会問題にもなっている。これを受けて、経済産業省では開発
プロセスのモデル策定や、ITスキル標準を規定したりする事で、改善に
取り組んできた。これらの実践論に基づく業界全体での開発の標準化は、
徐々に問題の解決に貢献しているものの、やはり、実際の問題解決とし
てはこれら標準化を踏まえた、より実践的なプロジェクトマネージメン
トスキルが必要とされる。
当カリキュラムでは、その習得方法として、
１．上流工程に当たる用件定義・設計から最終工程になる運用テストま
での工程を一貫して経験し、決められた制限のもと、要求を到達するた
めの方法論を構築し、各プロセスにおけるチェックレビューを行ないな
がら、開発プロジェクトを実践する。最初に改善目標を設定し、終了時
に目標に対する到達度評価を行なう。
２．開発するシステムとして、ソフトウエア会社の中で実際に使用する
基幹業務支援アプリケーションの構築を行なう。システムに関しては、
カリキュラム内にある要求定義により、後に受講者と決定する。通常の
業務支援システムではなく、コンポーネントの再利用と促進するための
基盤システムの構築のように、課題解決型・戦略投資型システムを構築
や、受注・請求管理のようなソフトウエア会社の全体像が展望できるシ
ステムの構築を行なう。
３．コーチや生徒同士が、ロールプレイをしながら、ユーザ側や開発側
になりきり双方の立場を経験しながら、開発プロセスを体験していく。
当カリキュラムでは、業界標準的プロセス管理手法、スキル評価手法を
学び、その後に、実際のソフトウエア会社での総括的システム、ないし
は、戦略的システムを、企画からテストまで構築し、当初策定した改善
目標に対しての到達度合いを評価することで、規範となるプロジェクト
マネージャの育成する事を目的にする。
# 工程 期間 納品物
1 要求仕様 2ヶ月 要求仕様書
2 基本設計 1ヶ月 基本設計書
3 詳細設計 3ヶ月 詳細設計書
4 製造 5ヶ月 ソフトウェア
5 単体テスト 2ヶ月 単体テスト項目票及び成績表
6 結合テスト 2ヶ月 詳細テスト項目票及び成績表
7 システムテスト 1ヶ月 システムテスト項目票及び成績表
8 運用テスト 1ヶ月 運用テスト項目票及び成績表
関連する学部の講義（専修科目ではありません）
a. ソフトウエア工学Ⅰ
b. ソフトウエア工学Ⅱ
c. ウェブプログラミング
・オブジェクト指向
・ウェブアプリケーションの配備・運用
履修を勧める科目
d. Requirements Engineering
e. Software project management
f. Introduction to Advanced Internet Technology and Application
[ソフトウエア工学論]
会社での実戦経験があれば良いです。学生においてはソフトウエア工学Iの履修で構いません。
[JAVA、.NET等の開発技術]
上記いずれかの言語での開発を予定しています。言語習得が必要となります。大学での該当するカリキュラム履修、職場での経験で構いません。
[データベース]
技術習得されている事を推薦しますが、経験がなくとも可能です。必要に応じて関連する科目を履修してもらいます。
チームでの開発作業と提出された成果物にもとづき、総合的に評価します。
The role of the Internet as a tool to get knowledge and valuable information on different topics has been increased drastically over the past years. Studying and working with tools to access the information and knowledge on the web are the important part of software engineering education. Search engines allow users to retrieve information from the Web using keywords and phrases. Many words and phrases have multiple meaning. This is a big problem for search engines. Search engines provide the users with small snippets of texts taken from the retrieved documents. Users usually cannot evaluate the relevancy of a document using this information. One promising way to improve the quality of retrieval is to design summarizing methods and applications for Web-browsers which can run on client computers or on servers of Internet service ptoviders. They will create automatically informative summaries for retrieved documents, and then present them to the end user. The aim is to reduce the amount of information to analyze manually.
We propose the National Institute of Informatics, Tokyo, Japan and Saint Petersburg State University, Russia as project supporters. NII is a world leading organization in Information Access technologies applied to Asia languages. This institute is our partner in research related to information retrieval. Saint Petersburg State University is a famous university with excellent education traditions in mathematics, programming and software engineering. This university is our partner in the project on text mining. Both institutions can provide support via consultations by key researcher and practitioners, round table discussions, etc. These communications can be done via the Internet using modern communication technologies. NII can provide their data collections they use in NTCIR Workshops which are major scientific forums for researchers in the information retrieval area.
Designing the architecture of the application software running according to the following scenario: The application takes the user query, submits it to the search engine, gets its response, downloads top 10 documents, creates informative summaries, and presents results to the end user. Aim is to get the necessary knowledge in designing the architecture of software systems.
a. Literature Study. Aim is to become familiar with the area of research and software development and to get the analytical skills necessary for any researcher and software development: This application area is very dynamic.
b. Investigating software market. Aim is to find appropriate software components applicable to the area of research.
c. Implementing the classical summarizing methods and test them utilizing data collections created by NII. Aim is to see the base level of the quality of summaries produced automatically.
d. Designing advanced methods on the base the latest results published in the annual conferences such as TREC, NTCIR Workshop, SIGIR, WWW, etc. Aim is to design software components on the base proposed ideas.
e. Conducting a set of tests to compare the quality of summaries designed applying classical and proposed methods. Aim is to select the most powerful method producing the best results.
f. Designing the architecture of the application software running according to the following scenario: The application takes the user query, submits it to the search engine, gets its response, downloads top 10 documents, creates informative summaries, and presents results to the end user. Aim is to get the necessary knowledge in designing the architecture of software systems.
g. Programming the prototype of the system (creating code). Aim is to get skills to work in a team for participants of this project.
h. Testing the prototype. Aim is to get knowledge about the full cycle of designing application software.
On-line documents will be used.
Prerequisites: Students have to have programming skills, be able to analyze the English scientific literature.
Benefits: After implementation of this project, students obtain the necessary skills to design, implement, and test the software system. They will get experience in designing WEB application software. This application area is growing very fast and the demand to design new software is very high.
The final grade will be calculated based on the following weights:
Presentations at group seminars ? 50%
Software development ? 50%
http://web-int.u-aizu.ac.jp/~vkluev/courses/SEA/