| 中地 祥剛 | M, 2回目発表 | ソフトウェア設計学 | 飯田 元, 井上 美智子, 市川 昊平, 崔 恩瀞 |
|
title: Proposal of request acquisition support environment for web application development using test case
Miscommunication between planners (e.g. people who provide request but cannot write/read code) and developers frequently occurs in the web application development process. To solve this problem, I propose a system that supports requirement elicitation process for the web application. The proposed system provides a feature that allows planners to write test cases using a simple domain specific language. It also converts input test cases into test code so that programmers can confirm the requirement by checking the test code. In this presentation, I will introduce the proposed system and my current progress for the master thesis. language of the presentation: Japanese | |||
| 鍛治 秀伍 | M, 2回目発表 | 情報セキュリティ工学 | 林 優一, 井上 美智子, 岡田 実, 藤本 大介 |
|
title: Study on Data Injection against Electronic Device with Intentional Weakened Immunity
abstract: The threats of data injection against electronic devices due to intentional electromagnetic interference using low-power electromagnetic waves had been targeted to devices that have potentially low electromagnetic tolerance (immunity) to interference waves. In contrast, when intentionally degrading the immunity of the device using a slight circuit modification, there is a possibility to expand the threat scope of the conventional data injection threats using low-power IEMI. In this study, the communication line between devices is intentionally circuit modified, and the electric signal represented arbitrary data is generated by irradiating low-power electromagnetic waves. Then, I show that it is possible to let the IC of the target device process the generated data as a legitimate signal. Also, I will consider the countermeasure against such attack from the viewpoint of electromagnetic compatibility. language of the presentation: Japanese | |||
| 任 翔太 | M, 2回目発表 | 情報セキュリティ工学 | 林 優一, 井上 美智子, 岡田 実, 藤本 大介 |
|
title: HT-detection method based on impedance measurements of ICs
abstract: Threats have been reported that Hardware Trojan (HT) is installed by attackers, which cause operational failure and/or information leakage. Conventional detection methods for HT mainly focus on the HT installed inside an integrated circuit (IC); thus, it is difficult to apply conventional method to this kind of threats immediately. Based on the above background, we propose a detection method for such HT implemented outside the IC by measuring electrical responses changing in the impedance of the wiring caused by HT implementation in time domain. language of the presentation: Japanese | |||
| 柴田 敦也 | M, 2回目発表 | 知能コミュニケーション | 中村 哲, 松本 健一, 須藤 克仁 |
|
title: Efficient algorithm suggesting system by using source-code on Online Judge System(OJS)
abstract: When we develop a software, we sometimes need a knowledge about algorithms. In this case, algorithm means like a shortest-path algorithm, bipartite matching algorithm and etc. For example, network system includes shortest-path problem when decide rooting part. Actually, Apache’s project also includes many algorithms. Additionally, in some cases, there is more than one algorithms to solve a problem. And each algorithm’s speck like a time complexity isn’t same. For example, in shortest-path problem, you can use below some algorithms Dijkstra’s algorithm, Bellman-ford algorithms and Floyd-Warshall algorithm. The algorithms which solves same problem can be separated by problem. For example, in many cases, Dijkstra’s algorithm is more efficient than bellman-ford algorithm. Therefore, choosing proper algorithms is necessary but difficult. Proper algorithm suggesting system from user’s source-code is useful, I think. In this study, I offer a system which suggests proper algorithms implementation sample to user by using OJS's source-codes. language of the presentation: Japanese | |||
| 浅野 幸之助 | M, 2回目発表 | 計算システムズ生物学 | 金谷 重彦, 松本 健一, 小野 直亮, 黄 銘, MD.ALTAF-UL-AMIN |
|
title: Feature extraction and Clustering analysis of Pancreatic Pathological Images Based on Unsupervised Convolutional Neural Network.
abstract: In recent years, the demand for Computer Aided Diagnosis(CAD) using Deep Learning is increasing on clinical site since there is a shortage of pathologists. Most studies of medical image processing based on supervised learning based on heuristically designed models that need appropriate labels according to the task, the prediction from the model changes depending on how to label it. At the clinical site, it is desirable to classify not only whether cancer exists or not, but also to evaluate the types and stages of the tumor in order to decide appropriate treatment plan. Pancreatic cancer is known as one of the poorerst prognosis cancer, due to the difficulty of early detection, the fast progression stages and the frequency of distant metastasis, etc. Estimation of medical features of tumor cell, such as the proliferation rate, metastatic state, etc. will provide much advance to lead to better treatment. In our study, we analyse the pancreatic histopathological image using Convolutional Neural Network (CNN) that trained two different dying methods to predict discrete representation and malignancy of the tissue. language of the presentation: Japanese | |||
| 武末 祥太郎 | M, 2回目発表 | 計算システムズ生物学 | 金谷 重彦, 松本 健一, 小野 直亮, 黄 銘, MD.ALTAF-UL-AMIN |
|
title:Regionalization analysis based on food supply across the Globe abstract: In the global food system, accessibility of food and region has been developed with being closely related. Identifying relationship between region and accessibility of foods plays a significant role in food security. In this study, we focus on the relationship between the food supply and global countries.We assess the regionalization of countries by hierarchical clustering method using food supply data, reveal characteristic geographic zones based on food supply. And we also examine climatic factors on these zones. language of the presentation:Japanese | |||