| | SLO | ENG | Piškotki in zasebnost

Večja pisava | Manjša pisava

Iskanje po katalogu digitalne knjižnice Pomoč

Iskalni niz: išči po
išči po
išči po
išči po
* po starem in bolonjskem študiju

Opcije:
  Ponastavi


1 - 3 / 3
Na začetekNa prejšnjo stran1Na naslednjo stranNa konec
1.
A VAN-Based Multi-Scale Cross-Attention Mechanism for Skin Lesion Segmentation Network
Shuang Liu, Zeng Zhuang, Yanfeng Zheng, Simon Kolmanič, 2023, izvirni znanstveni članek

Opis: With the rise of deep learning technology, the field of medical image segmentation has undergone rapid development. In recent years, convolutional neural networks (CNNs) have brought many achievements and become the consensus in medical image segmentation tasks. Although many neural networks based on U-shaped structures and methods, such as skip connections have achieved excellent results in medical image segmentation tasks, the properties of convolutional operations limit their ability to effectively learn local and global features. To address this problem, the Transformer from the field of natural language processing (NLP) was introduced to the image segmentation field. Various Transformer-based networks have shown significant performance advantages over mainstream neural networks in different visual tasks, demonstrating the huge potential of Transformers in the field of image segmentation. However, Transformers were originally designed for NLP and ignore the multidimensional nature of images. In the process of operation, they may destroy the 2D structure of the image and cannot effectively capture low-level features. Therefore, we propose a new multi-scale cross-attention method called M-VAN Unet, which is designed based on the Visual Attention Network (VAN) and can effectively learn local and global features. We propose two attention mechanisms, namely MSC-Attention and LKA-Cross-Attention, for capturing low-level features and promoting global information interaction. MSC-Attention is designed for multi-scale channel attention, while LKA-Cross-Attention is a cross-attention mechanism based on the large kernel attention (LKA). Extensive experiments show that our method outperforms current mainstream methods in evaluation metrics such as Dice coefficient and Hausdorff 95 coefficient.
Ključne besede: CNNs, deep learning, medical image processing, NLP, semantic segmentation
Objavljeno v DKUM: 14.03.2024; Ogledov: 496; Prenosov: 311
.pdf Celotno besedilo (1,46 MB)
Gradivo ima več datotek! Več...

2.
Reinforced piled embankment for a high-speed railway over soft soil : a numerical and analytical investigation
Yan Zhuang, Xiaoyan Cui, 2015, izvirni znanstveni članek

Opis: A geosynthetic, reinforced, piled embankment is an effective and economic method to solve the problems of possible bearing failure, unacceptable settlement and slope instability for an embankment built over soft soil; this has led to its widespread use, especially for high-speed railway embankments. Some design methods have been developed to assess the performance of these reinforced structures, which are mainly based on the results from small-scale models and numerical simulations. However, the reliability of these methods needs to be validated under full-scale field tests. This paper presents a numerical and analytical study for a full-scale field test of the Fengyang high-speed railway embankment. The results were analyzed and discussed in terms of the settlement of subsoil, the stressconcentration ratio (SCR), the axial force and the frictional stress of the pile. They showed that the settlement of the subsoil, from both the finite-element method (FEM) and the analytical method, were in good agreement with the measurement, and thus was a reliable parameter to assess the performance of the piled embankment with reasonable accuracy. The SCR was overestimated by the modified Terzaghi method, with a difference of 25%, while it was underestimated by the FEM, with a difference of approximately 20%. It was also shown that the tensile force in the reinforcement could be effectively assessed using the proposed analytical method, while it was overestimated by the FEM with a difference of 44%.
Ključne besede: reinforced piled embankment, high-speed railway, numerical simulation, analytical method
Objavljeno v DKUM: 15.06.2018; Ogledov: 1746; Prenosov: 241
.pdf Celotno besedilo (578,98 KB)
Gradivo ima več datotek! Več...

3.
A biometric authentication model using hand gesture images
Simon Fong, Yan Zhuang, Iztok Fister, Iztok Fister, 2013, izvirni znanstveni članek

Opis: A novel hand biometric authentication method based on measurements of the user's stationary hand gesture of hand sign language is proposed. The measurement of hand gestures could be sequentially acquired by a low-cost video camera. There could possibly be another level of contextual information,associated with these hand signs to be used in biometric authentication. As an analogue, instead of typing a password 'iloveu' in text which is relatively vulnerable over a communication network, a signer can encode a biometric password using a sequence of hand signs, 'i', 'l', 'o', 'v', 'e', and 'u'. Subsequently the features from the hand gesture images are extracted which are integrally fuzzy in nature, to be recognized by a classification model for telling if this signer is who he claimed himself to be, by examining over his hand shape and the postures in doing those signs. Itis believed that everybody has certain slight but unique behavioral characteristics in sign language, so are the different hand shape compositions. Simple and efficient image processing algorithms are used in hand sign recognition, including intensity profiling, color histogram and dimensionality analysis, coupled with several popular machine learning algorithms. Computer simulation is conducted for investigating the efficacy ofthis novel biometric authentication model which shows up to 93.75% recognition accuracy.
Ključne besede: biometric authentication, hand gesture, hand sign recognition, machine learning
Objavljeno v DKUM: 28.06.2017; Ogledov: 1682; Prenosov: 467
.pdf Celotno besedilo (1,83 MB)
Gradivo ima več datotek! Več...

Iskanje izvedeno v 0.04 sek.
Na vrh
Logotipi partnerjev Univerza v Mariboru Univerza v Ljubljani Univerza na Primorskem Univerza v Novi Gorici