Exploring and annotating collections of images without meta-data is a complex task which requires convenient ways of presenting datasets to a user. Visual analytics and information visualization can help users by prov...
详细信息
ISBN:
(纸本)9781450347013
Exploring and annotating collections of images without meta-data is a complex task which requires convenient ways of presenting datasets to a user. Visual analytics and information visualization can help users by providing interfaces, and in this paper, we present an open source application that allows users from any domain to use feature-based clustering of large image collections to perform explorative browsing and annotation. For this, we use various image feature extraction mechanisms, different unsupervised clustering algorithms and hierarchical image collection visualization. The performance of the presented open source software allows users to process and display thousands of images at the same time by utilizing heterogeneous resources such as GPUs and different optimizationtechniques.
Feature visualization is one of the most popular techniques used to interpret the internal behavior of individual units of trained deep neural networks. Based on activation maximization, they consist of finding synthe...
详细信息
ISBN:
(纸本)1577358872
Feature visualization is one of the most popular techniques used to interpret the internal behavior of individual units of trained deep neural networks. Based on activation maximization, they consist of finding synthetic or natural inputs that maximize neuron activations. This paper introduces an optimization framework that aims to deceive feature visualization through adversarial model manipulation. It consists of fine-tuning a pre-trained model with a specifically introduced loss that aims to maintain model performance, while also significantly changing feature visualization. We provide evidence of the success of this manipulation on several pre-trained models for the classification task with ImageNet.
In this paper we use computational intelligence numerical optimizationtechniques to design optimum finite impulse response (FIR) digital filters. Results will be presented and compared with three different numerical ...
详细信息
A new parallel volume rendering method for large unstructured data has been developed for the Earth Simulator. Concurrent visualization can be performed with numerical computation modules on the supercomputer. The sup...
详细信息
ISBN:
(纸本)0889864187
A new parallel volume rendering method for large unstructured data has been developed for the Earth Simulator. Concurrent visualization can be performed with numerical computation modules on the supercomputer. The supervoxel partition and voxel resampling techniques were adopted to get high parallel performance for extremely large complicated unstructured datasets. In order to reduce the number of voxels and keep enough precision in the resampling method, a detail-preserving voxel resampling method is presented in which some original unstructured grid elements are added into important voxels. Four kinds of details are defined in the method, including volume, interval, boundary and isosurface details. The implementation and optimization strategies for the Earth Simulator are described according to its hardware architecture. The experimental results show the feasibility and effectiveness of the proposed method.
Datasets of tens of gigabytes are becoming common in computational and experimental science. Providing remote visualization of these large datasets with adequate levels of quality and interactivity is an extremely cha...
详细信息
ISBN:
(纸本)9780819469533
Datasets of tens of gigabytes are becoming common in computational and experimental science. Providing remote visualization of these large datasets with adequate levels of quality and interactivity is an extremely challenging task, particularly for scientists who collaborate in widely distributed locations and their primary access to visualization resources is a desktop computer. This paper describes a remote visualization system for large-scale terrain rendering based on parallel streaming pipeline architecture. The visualization pipeline is divided in a client-server paradigm to take advantage of the powerful computing and storage resources on the dedicated computers. The two key components of this framework are: view-dependent simplification of the terrain mesh;and a scheme for delivering a minimally necessary subset of triangle strips to any user requesting an interactive visualization session. To verify the effectiveness of proposed schemes and data structures, the prototype system was implemented on China next-generation Internet backbone. Approximate 60GB size image resources for flight simulation were stored centrally in Wuhan, whereas scientists geographically dispersed in Beijing and Shanghai could manipulate and visualize these large 3D datasets in an efficient and flexible way, furthermore, the need for data replication to local desktops was eliminated.
We present a novel approach for fast and accurate collision detection using programmable graphics processing unit (GPU). The approach maps the collision detection between two triangular meshes with arbitrary shapes in...
详细信息
ISBN:
(纸本)0889864543
We present a novel approach for fast and accurate collision detection using programmable graphics processing unit (GPU). The approach maps the collision detection between two triangular meshes with arbitrary shapes into GPU based streaming computations using floating point buffer. Doing, so. the collision detection problem becomes to use programmable GPU for finding the intersections between a collection of line segments and a set of triangles. Occlusion queries are then employed to inquire about the collision detection result. Several optimizationtechniques are developed to further improve the efficiency of collision detection. Our approach has been implemented on commodity PC and experimental results are promising.
In the recent scenarios, Computational Intelligence techniques are widely applied to solve complex scientific and mathematical problems which involve variety and huge volume of data. The characteristics of data have g...
详细信息
ISBN:
(纸本)9781509049974
In the recent scenarios, Computational Intelligence techniques are widely applied to solve complex scientific and mathematical problems which involve variety and huge volume of data. The characteristics of data have great influence on the behaviour of Nature Inspired Computing (NIC) algorithms to find optimal or near optimal solution while solving non-linear complicated problems. In this paper, the performance of the newly developed Synergistic Fibroblast optimization (SFO) algorithm on solving different real world problems which encompass diverse sort of dataset has been investigated and demonstrated its effectiveness. The significant outcomes have revealed that the novel SFO algorithm is efficient and compatible with various types of data, computational methods and techniques to produce promising results in both qualitative and quantitative perspectives.
The projection of high-dimensional data by linear or non-linear techniques is a well established technique in pattern recognition and other scientific and industrial application fields. Commonly, methods affiliated to...
详细信息
ISBN:
(纸本)0769522912
The projection of high-dimensional data by linear or non-linear techniques is a well established technique in pattern recognition and other scientific and industrial application fields. Commonly, methods affiliated to multidimensional-scaling, projection pursuit or Sammons non-linear distance preserving mapping are applied, based on gradient descent techniques. These suffer from well known dependence on initial or starting value and their limited ability to reach only local minimum. In this paper stochastic search techniques are applied to the NLM to achieve lower residual stress or error value in competitive time. Encouraging results have been obtained for a particular developed local algorithm both with regard to convergence time and residual error
Policy making is a very complex task taking into account several aspects related to sustainability, namely impact on the environments, health of productive sectors, economic implications and social acceptance. Optimiz...
详细信息
ISBN:
(纸本)9783642406263;9783642406270
Policy making is a very complex task taking into account several aspects related to sustainability, namely impact on the environments, health of productive sectors, economic implications and social acceptance. optimization methods could be extremely useful for analysing alternative policy scenarios, but should be complemented with several other techniques such as machine learning, agent-based simulation, opinion mining and visualization to come up with an integrated system able to support decision making in the overall policy design life cycle. I will discuss how these techniques could be merged with optimization and I will identity some open research directions.
The convex feasibility problem of finding a point in the intersection of finitely many nonempty closed convex sets in the Euclidean space has many applications in various fields of science and technology, particularly...
详细信息
ISBN:
(纸本)081944281X
The convex feasibility problem of finding a point in the intersection of finitely many nonempty closed convex sets in the Euclidean space has many applications in various fields of science and technology, particularly in problems of image reconstruction from projections, in solving the fully discretized inevrse problem in radiation therapy treatment planning, and in other image processing problems. Solving systems of linear equalities and/or inequalities is one of them. Many of the existing algorithms use projections onto the sets and may: (i) employ orthogonal-, entropy-, or other Bregman-projections, (ii) be structurally sequential, parallel, block-iterative, or of the string-averaging type, (iii) asymptotically converge when the underlying system is, or is not, consistent, (iv) solve the convex feasibility problem or find the projection of a given point onto the intersection of the convex sets, (v) have good initial behavior patterns when some of their parameters are appropriately chosen.
暂无评论