Zobrazeno 1 - 10
of 1 107
pro vyhledávání: '"P Töws"'
This paper proposes a novel hue-like angular parameter to model the structure of deep convolutional neural network (CNN) activation space, referred to as the {\em activation hue}, for the purpose of regularizing models for more effective learning. Th
Externí odkaz:
http://arxiv.org/abs/2310.03911
Downsampling layers, including pooling and strided convolutions, are crucial components of the convolutional neural network architecture that determine both the granularity/scale of image feature analysis as well as the receptive field size of a give
Externí odkaz:
http://arxiv.org/abs/2306.11982
Autor:
Manashti, Javad, Pirnia, Pouyan, Manashty, Alireza, Ujan, Sahar, Toews, Matthew, Duhaime, François
This project aimed to determine the grain size distribution of granular materials from images using convolutional neural networks. The application of ConvNet and pretrained ConvNet models, including AlexNet, SqueezeNet, GoogLeNet, InceptionV3, DenseN
Externí odkaz:
http://arxiv.org/abs/2303.04269
This study aims to evaluate PSDNet, a series of convolutional neural networks (ConvNets) trained with photographs to predict the particle size distribution of granular materials. Nine traditional feature extraction methods and 15 pretrained ConvNets
Externí odkaz:
http://arxiv.org/abs/2303.04265
We report on a novel model linking deep convolutional neural networks (CNN) to biological vision and fundamental particle physics. Information propagation in a CNN is modeled via an analogy to an optical system, where information is concentrated near
Externí odkaz:
http://arxiv.org/abs/2206.02220
This paper proposes to extend local image features in 3D to include invariance to discrete symmetry including inversion of spatial axes and image contrast. A binary feature sign $s \in \{-1,+1\}$ is defined as the sign of the Laplacian operator $\nab
Externí odkaz:
http://arxiv.org/abs/2205.15456
Autor:
Carluer, Jean-Baptiste, Chauvin, Laurent, Luo, Jie, Wells III, William M., Machado, Ines, Harmouche, Rola, Toews, Matthew
This work details a highly efficient implementation of the 3D scale-invariant feature transform (SIFT) algorithm, for the purpose of machine learning from large sets of volumetric medical image data. The primary operations of the 3D SIFT code are imp
Externí odkaz:
http://arxiv.org/abs/2112.10258
Autor:
Chauvin, Laurent, Toews, Matthew
Subject ID labels are unique, anonymized codes that can be used to group all images of a subject while maintaining anonymity. ID errors may be inadvertently introduced manually error during enrollment and may lead to systematic error into machine lea
Externí odkaz:
http://arxiv.org/abs/2110.04055
Autor:
Elodie Wielgus, Maik Henrich, Christian Fiderer, Ariana Töws, Jan‐Niklas Michel, Franz Kronthaler, Marco Heurich
Publikováno v:
Ecological Solutions and Evidence, Vol 5, Iss 2, Pp n/a-n/a (2024)
Abstract Human activities can affect the behaviour and fitness of wildlife. However, the response of animals to nonlethal human activities has not been well‐studied in wild boar, Sus scrofa, even though it is a widespread species in Europe and has
Externí odkaz:
https://doaj.org/article/170a0688d7d24308a323d280cbf845c6
We propose a novel pairwise distance measure between image keypoint sets, for the purpose of large-scale medical image indexing. Our measure generalizes the Jaccard index to account for soft set equivalence (SSE) between keypoint elements, via an ada
Externí odkaz:
http://arxiv.org/abs/2103.06966