【速報】IEEE WCCI 2012

IEEE WCCI 2012がブリスベン・オーストラリアで開催されます。

Friday, IEEE CEC, FrC 4-2, 13:30-14:30, Applications of Evolutionary Computation In Biomedical Engineering, Jose A. Lozano
535, Hiroaki Yamaguchi, Tomoyuki Hiroyasu, Sakito Nunokawa, Noriko Koizumi, Naoki Okumura, Hisatake Yokouchi, Mitsunori Miki and Masato Yoshimi, Comparison Study of Controlling



報告者氏名 山口 浩明
発表論文タイトル 細胞画像分割問題のためのGPにおけるブロート抑制モデルの比較
発表論文英タイトル Comparison Study of Controlling Bloat Model of GP in Constructing Filter for Cell Image Segmentation Problems
著者 山口浩明,廣安知之,布川将来人,小泉範子,奥村直穀,横内久猛,三木光範,吉見真聡
講演会名 2012 IEEE World Congress on Computational Intelligence
会場 Brisbane Convention and Exhibition Centre, Australia
開催日程 2012/06/10-2012/06/15

1. 講演会の詳細
2012/06/10から2012/06/15にかけて,ブリスベン(オーストラリア)にて開催されました2012 IEEE World Congress on Computational Intelligence (WCCI)に参加いたしました.この会議は,2年に1度IEEEによって主催され,ニューラルネットワークをテーマにしたIJCNN,ファジィのFUZZ-IEEE,進化計算のCECの3つの会議が同時に行われます.今回,私はCECにおいて口頭発表を行いました.また本研究室からは他に廣安先生が参加しました.
2. 研究発表
2.1. 発表概要
私は15日の13:30からのセッション「Applications of Evolutionary Computation In Biomedical Engineering」に参加いたしました.発表の形式は口頭発表で,15分の講演時間と3分の質疑応答時間となっておりました.

本研究では,角膜再生医療を支援するための細胞画像解析システムの構築を目指している.既存の画像解析ソフトは,解析の際,使用者が画像処理を自ら組み合わせる必要があるため画像処理に関する知識が必要となる.そこで,遺伝的プログラミング(Genetic Programming:GP)を用い自動で目的の画像処理を構築できる方法が数多く提案されている.しかし,それらの手法では標準的なGPモデルの適用しか行われていない.そこで,本稿では,提案されているGPのブロート抑制モデルを適用し,細胞領域分割のために適したGPモデルを調査した.適用したモデルは,標準的なモデルに加え,Double Tournament,Tarpeian,Non-Destructive Crossover (NDC),Recombinative Hill-Clibming (RHC),Spatial Structure + Elitism (SS+E)の6種類である.これらのモデルに対して角膜内皮細胞画像を対象に,GPによって求められた画像処理の組み合わせと,そのロバスト性能の比較実験を行った.実験結果では,SS+Eモデルが木の深さ制限やペナルティのパラメータに依存せず,生成された細胞領域分割のための画像処理,ロバスト性能の両方において他のモデルよりも優れた値を示した.

2.2. 質疑応答
質問内容は,「今回比較したSS+Eモデルでの子個体の生成方法. 親と子の適合度が等しい場合はどういう処理をするか.」でした.1つ目の質問には,子個体一つ一つはそれぞれのサブ母集団内で生成されると回答しました.2つ目の質問は,その場で内容を理解できず,回答できませんでした.
2.3. 感想
3. 聴講

発表タイトル       : Image Description Generation without Image Processing using Fuzzy Inference
著者                  : Naho Ito and Masafumi Hagiwara
セッション名       : Fuzu IEEE Poster
Abstruct            : We propose a sentence generation method that describes images. We do not use image processing technique in our proposed method. Human annotated image tags are used as image information to generate sentence. By using human annotated tags, we think this enables to describe image more relevant and user specific. Our method uses Kyoto University’s case frame data and Google N-gram to generate candidate sentences. We extend these candidates to describe images more relevant. To be more precise, we added segments with missing semantic role, and added modification segments. To select one output sentence, we used fuzzy rules to grade naturalness of candidate sentences. To grading image relevance of the sentence, we scored word similarity for each word. The performance of the proposed system has been evaluated by subjective experiments and obtained satisfactory results.


発表タイトル       : Estimating Subjective Assessments using a Simple Biosignal Sensor
著者                  : Yoshihito Maki, Genma Sano, Yusuke Kobashi, Tsuyoshi Nakamura, Masayoshi Kanoh, Koji Yamada
セッション名       : FUZZ IEEE, SS human Symbiotic Systems
Abstruct            : Given the remarkable recent progress in robotics re- search, we can envision the day when robots and humans coexist and robots become closely integrated into our daily lives. This means endowing robots with the ability to communicate so they perceive human emotion, adapt their behavior to humans, and sense situations even without explicit instructions. Meanwhile, affective computing, that interprets emotion or other affective phenomena from human biosignals, has emerged as an area of great interest. In addition to biosignals-brain waves, heart rate, pulse, electrical activity, and the like-affective computing is concerned with facial expressions, gestures, and a wide range of other indicators of emotion. Here we explore the latest insights of affective computing in relation to human-robot interaction (HRI). There is good reason to believe robots will soon have the ability to read human emotions, so here we investigate the feasibility of inferring human psychological states from biosensor signals. Obviously, non-invasive biosensors that don’t interfere with normal everyday activities would be preferable. A number of inexpensive user-friendly brain-wave sensors have been brought to market recently, and we employ one of these devices, the NeuroSky Mindset EEG neuroheadset, in assessment trials to explore the feasibility of inferring subjective assessments. Using our experimental setup, we find that it is indeed possible to infer subjective assessments from biosignals, and this capability could prove immensely useful for future HRI applications.

この発表は,生体センサを用い人間の感情を推定するといった内容でした.生体センサにはNeuro Sky2)と呼ばれるヘッドフォン型の脳波測定装置を用いており,私たちの研究室と似たような研究でした.英語が聞き取れず,詳しい内容については残念ながらわからなかったのですが,Neuro SkyはNIRS以上に低拘束性であるため,より普段の生活に近い行動を計測するのが可能になると感じました.

発表タイトル       : Probabilistic Model Building GP with Belief Propagation
著者                  : Hiroyuki Sato, Yoshihiko Hasegaa, Danushaka Bollegala, Hitoshi Iba
セッション名       : IEEE CEC, Estimation of distribution algorithms
Abstruct            : Estimation of distribution algorithms (EDAs) which deal with tree structures as GP are called as probabilistic model building GPs (PMBGPs), and they show better search performance than GP in many problems. A problem of prototype tree-based method, a type of PMBGPs, is that samplings do not always generate the most probable solution, which is the individual with the highest probability and reflects a learned distribution most. This problem wastes a part of learning and increases the number of evaluations to get an optimum solution. In order to overcome this difficulty, this paper proposes a hybrid approach using Belief propagation (BP) in sampling process. BP is an inference algorithm on graphical models and can generate the most probable solution. By applying our approach to benchmark tests, we show that the proposed method is more effective than PLS alone.

この発表は,GPの確率モデル手法の一つであるPOLEに対して,Belief propagationと呼ばれる伝搬モデルを使用することを提案しています.対象問題はわからなかったのですが,提案手法がSGPや従来のPOLEよりも優れた性能を示しており,今後確率モデルについて私の研究に取り入れようと考えているので,再度深く調査したく思いました.

発表タイトル       : Brain Signal Pattern of Engrossed Subjects using Near Infrared Spectroscopy (NIRS) and its Application to TV Commercial Evaluation
著者                  : Mototaka Yoshioka, Tsuyoshi Inoue and Jun Ozawa
セッション名       : IEEE IJCNN, Brain Machines Interfaces
Abstruct            : In this paper, we present near infrared spectroscopy (NIRS) signal patterns of subjects when they are focused on specific tasks. We determined that oxygenated hemoglobin in the frontal cortex decreased when the subjects were engrossed in tasks, and we propose an evaluation method for TV commercials based on the results. TV commercials are produced to be as attractive as possible and can increase consumer awareness of a particular product or its features. Our idea is based on the assumption that the attractiveness of a commercial can be estimated by the extent of decrease in oxygenated hemoglobin using NIRS, and the consumer’s awareness of the product’s features in TV commercials can be measured by analyzing the subject’s glancing regions using an eye-tracking system. We obtained good agreement between the correlation of awareness and focus, and the possibility of estimating these parameters using NIRS is suggested.


発表タイトル       : Adaptive Formation Behaviors of Multi-robot for Cooperative Exploration
著者                  : Yutaka Yasuda, Naoyuki Kubota and Yuichiro Toda
セッション名       : Hybrid, Computational Intelligence for Cognitive Robotics
Abstruct            : This paper proposes a method for constituting the formation of a multi-robot system according to dynamically changing environments. First, we apply a method of multi- objective behavior coordination for integrating behavior outputs from the fuzzy control for collision avoidance and target tracing. Second, we apply a spring model to calculate the temporary target position of each robot for the formation behavior. Third, we discuss multi-robot behaviors based on the concept of coupling. The tight coupling is realized by the spring model while the loose coupling is realized by the individual decision making based on connection and disconnection with other robots. Furthermore, the proposed method is applied to the exploration in unknown environments. Finally, we discuss the effectiveness of the proposed method through several simulation results.

この発表は,複数のロボットが動的に変化する環境に応じた操作のための手法を提案していました. ファジィやロボット工学の知識がないため,理解が難しかったのですが,本手法では疎結合だけでなく,密結合の場合においても良好な結果を示していました.結果はシミュレーションであったので,実環境での展望が考えられます.また,日本人学生の発表でしたが,質疑応答では苦労しながら自身で対応しており,良い発表でした.
1)    2012 IEEE WCCI,http://www.ieee-wcci2012.org/
2)    Neuro Sky, http://www.neurosky.jp/