Probability Distributions.- Linear Models for Regression.- Linear Models for Classification.- Neural Networks.- Kernel Methods.- Sparse Kernel Machines.- Graphical Models.- Mixture Models and EM.- Approximate Inference.- Sampling Methods.- Continuous Latent Variables.- Sequential Data.- Combining Models.

Pattern Recognition and Machine Learning

“Unpaired Image-to-Image Translation using Cycle-Consistent Adversarial Networks”の学習報告

From the Publisher: 
The accessible presentation of this book gives both a general view of the entire computer vision enterprise and also offers sufficient detail to be able to build useful applications. Users learn techniques that have proven to be useful by first-hand experience and a wide range of mathematical methods. A CD-ROM with every copy of the text contains source code for programming practice, color images, and illustrative movies. Comprehensive and up-to-date, this book includes essential topics that either reflect practical significance or are of theoretical importance. Topics are discussed in substantial and increasing depth. Application surveys describe numerous important application areas such as image based rendering and digital libraries. Many important algorithms broken down and illustrated in pseudo code. Appropriate for use by engineers as a comprehensive reference to the computer vision enterprise.

Computer vision : a modern approach = 计算机视觉 : 一种现代的方法

Color image segmentation: advances and prospects

When considering person re-identification (re-ID) as a retrieval process, re-ranking is a critical step to improve its accuracy. Yet in the re-ID community, limited effort has been devoted to re-ranking, especially those fully automatic, unsupervised solutions. In this paper, we propose a k-reciprocal encoding method to re-rank the re-ID results. Our hypothesis is that if a gallery image is similar to the probe in the k-reciprocal nearest neighbors, it is more likely to be a true match. Specifically, given an image, a k-reciprocal feature is calculated by encoding its k-reciprocal nearest neighbors into a single vector, which is used for re-ranking under the Jaccard distance. The final distance is computed as the combination of the original distance and the Jaccard distance. Our re-ranking method does not require any human interaction or any labeled data, so it is applicable to large-scale datasets. Experiments on the large-scale Market-1501, CUHK03, MARS, and PRW datasets confirm the effectiveness of our method.

/pdf/re-ranking-person-re-identification-with-k-reciprocal-3erzmilgv1.pdf

Re-ranking Person Re-identification with k-Reciprocal Encoding

This paper gives an overview of the Sensing Web project, launched in 2007 in Japan. The project's aim is to open the data obtained by the sensors existing in our daily living environment for various purposes. Since the data obtained by observing the real world directly with sensors include real-world information different from the Web, a new worldwide social information infrastructure - Sensing Web - is realized. In this article, we discuss the research issues for arising in connection with the Sensing Web.

Sensing web: to globally share sensory data avoiding privacy invasion

あらまし 視体積交差法では，対象物体を観測するカメラ台数が多ければ多いほど復元形状の再現性が向上するが，カメラを配置できる空間の制約によりカメラ台数を無制限に増やすことは難しい．これに対して，もし対象物体の剛体運動を各時刻の視体積から推定することができれば，複数の時刻でのカメラ画像を統合して，一つの時刻で得る視体積よりもより再現性の高い視体積を得ることができる．そこで我々は，この考え方に基づき，運動する剛体物体を観測することで，時系列で得られる対象物体の画像を統合し，多くのカメラを用いたときと同等の復元形状を少ないカメラ台数で得ることを目指す．この際に対象物体の運動を推定する必要があるが，これは特徴点を各時刻の視体積から抽出することで行う．特徴点抽出で問題となるのは，視体積が物体が占める領域以外に余分な領域を含むことである．このために，物体表面に現れる特徴的な部分を特徴点として抽出しようとすると，その特徴点が余分な領域に覆われて視体積表面上に現れないことが起こる．つまり，各時刻の視体積から時系列で安定して抽出される特徴点を得ることが難しいのである．我々は各時刻ごとのカメラ画像から視体積を求め，更に視体積をシルエットに投影し，余分な領域に覆われない点を抽出する手法を提案する．このような点を特徴点として用い，精度の良い運動推定を行うことができた．シミュレーション実験において，多くのカメラを用いたときと同等の復元形状を得られることを確認し，実画像を用いた実験でも復元形状の再現性の向上を確認した．キーワード 視体積交差法，特徴点抽出，運動推定，時系列データ，標本化誤差

An Accurate Shape Reconstruction Method by Integrating Visual Hulls in Time Sequences

Realistic simulation and manipulation of virtual cloth based on energy minimization techniques require higher computations. A model realizing force as well as shape to manipulate the virtual cloth demands even more computational cost. Mesh resolution of cloth has a great effect on the computational cost of cloth simulation. Therefore, granularity of mesh must be chosen in a manner that represents the cloth with a minimum computational cost and maintaining the quality (reality) of simulation. In this paper, we propose the adaptive mesh refinement combined with coarsening (AMRC) to have maximum possible efficiency and to maintain the acceptable reality. In addition, we propose an efficient parallel implementation in OpenMP to speed up the processing. AMRC changes the mesh density during execution time and requires the dynamic redistribution of load. A scheme of creating Active-List has been developed to manage the load balancing. We have parallelized the cloth simulation with a satisfactory efficiency by analyzing communication and synchronization overheads caused by OpenMP.

http://www.mm.media.kyoto-u.ac.jp/old/research/doc/575/420-079.pdf

Simulating realistic force and shape of virtual cloth with adaptive meshes and its parallel implementation in OpenMP.

False Alert Rejection for Detecting Objects on a Table by Touch Reasoning

We propose a method for speech and video indexing on automatic lecture recording system for archiving lectures. We previously proposed an automatic imaging method for lectures based on “Dynamic Situation”, which describes the situation in the lecture room. The imaging method can segment and index the video image based on the dynamic situation. The dynamic situation is estimated by the location and the speech volume of a lecturer and students. We apply the indexing method to the automatic archiving system for lectures. key words Lecture Recording, Archive, Multimedia, Video Indexing, Hypertext

Michihiko Minoh

Papers

Sensing web: to globally share sensory data avoiding privacy invasion

An Accurate Shape Reconstruction Method by Integrating Visual Hulls in Time Sequences

Simulating realistic force and shape of virtual cloth with adaptive meshes and its parallel implementation in OpenMP.

False Alert Rejection for Detecting Objects on a Table by Touch Reasoning

Speech and Video Indexing on Automatic Lecture Recording System