1,721,061 research outputs found
Comparing NLP based Strategies for Web Querying
No full textThis paper describes an experiment aiming at evaluating the role of NLP based optimizations (i.e. morphological derivation and synonymy expansion) in web search strategies. Keywords and their expansions are composed in two different Boolean expressions (i.e. expansion intertion and Cartesian combination) and then compared with a deyword conjunctive composition, considered as the baseline. Results confirm the hypothesis that linguistic optimizations significantly improve the search engine performance
A linear approach for sparse coding by a two-layer neural network
No full textMany approaches to transform classification problems from non-linear to linear by feature transformation have been recently presented in the literature. These notably include sparse coding methods and deep neural networks. However, many of these approaches require the repeated application of a learning process upon the presentation of unseen data input vectors, or else involve the use of large numbers of parameters and hyper-parameters, which must be chosen through cross-validation, thus increasing running time dramatically. In this paper, we propose and experimentally investigate a new approach for the purpose of overcoming limitations of both kinds. The proposed approach makes use of a linear auto-associative network (called SCNN) with just one hidden layer. The combination of this architecture with a specific error function to be minimized enables one to learn a linear encoder computing a sparse code which turns out to be as similar as possible to the sparse coding that one obtains by re-training the neural network. Importantly, the linearity of \{SCNN\} and the choice of the error function allow one to achieve reduced running time in the learning phase. The proposed architecture is evaluated on the basis of two standard machine learning tasks. Its performances are compared with those of recently proposed non-linear auto-associative neural networks. The overall results suggest that linear encoders can be profitably used to obtain sparse data representations in the context of machine learning problems, provided that an appropriate error function is used during the learning phase
Computer-aided Detection of Clustered Microcalcifications in Digitized Mammograms using Gabor Functions
No full text
An Action-tuned Neural Network Architecture for Hand Pose Estimation
No full textThere is a growing interest in developing computational models of grasping action recognition. This interest is increasingly motivated by a wide range of applications in robotics, neuroscience, HCI, motion capture and other research areas. In many cases, a vision-based approach to grasping action recognition appears to be more promising. For example, in HCI and robotic applications, such an approach often allows for simpler and more natural interaction. However, a vision-based approach to grasping action recognition is a challenging problem
due to the large number of hand self-occlusions which make the mapping from hand visual appearance to the hand pose an inverse ill-posed problem. The approach proposed here builds on the work of Santello
and co-workers which demonstrate a reduction in hand variability within a given class of grasping actions.
The proposed neural network architecture introduces specialized modules for each class of grasping actions and viewpoints, allowing for a more robust hand pose estimation. A quantitative analysis of the proposed
architecture obtained by working on a synthetic data set is presented and discussed as a basis for further work
Computer aided detection of clustered microcalcifications in digitized mammograms using Gabor functions
No full textThis paper presents a multiresolution approach to the computer aided detection of clustered microcalcifications in digitized mammograms based on Gabor elementary functions. A bank of Gabor functions with varying spatial extent and tuned to different spatial frequencies is used for the extraction of microcalcifications characteristics. Classification is performed by an Artificial Neural Network with supervised learning.First results show that most microcalcifications, isolated or clustered, are detected by our algorithm with a 95\% value both for sensibility and specificity as measured on a test data set
MEDIARAD: a Multiplatform Software Environment for Developing Imaging Applications in RADiology
No full textIn this paper we describe the principles and the design issues of MEDIARAD, a platform-independent user-oriented programming environment for developing imaging applications in Radiology. The development of such system is motivated by the significant increase, during the last few years, of the demand for computerized medical imaging systems in radiology. This increase is due mainly to the advent of newer imaging modalities, such as magnetic resonance or computerized tomography, as well as to the activation of several radiological Screening Programs for early diagnosis of cancer in most western Countries. MEDIARAD should be useful at least to four different types of users:1) physicians and radiologists who are the basic users and simply want a Computer-Aided Detection (CAD) system in order to receive help in the diagnostic process; 2) Computer Vision researchers and software developers who look for suitable tools to easily and effortless build their own CAD applications; 3) computer trained physicians, who might want to interact with the system in order to personalize and improve it, and 4) researchers interested mainly in testing specific algorithms during the development and evaluation stages which lead to building a specific imaging application. MEDIARAD has already been used to build a CAD system for detecting clusters of breast microcalcifications in digitized mammograms
- …
