Results 41 to 50 of about 1,673,879 (163)

Combining Sentiment Lexica with a Multi-View Variational Autoencoder [PDF]

, 2019
When assigning quantitative labels to a dataset, different methodologies may rely on different scales. In particular, when assigning polarities to words in a sentiment lexicon, annotators may use binary, categorical, or continuous labels.
Augenstein, Isabelle, Cotterell, Ryan, Hoyle, Alexander, Wallach, Hanna, Wolf-Sonkin, Lawrence   +4 more
core   +2 more sources

Multi-view 3D retrieval using silhouette intersection and multi-scale contour representation [PDF]

, 2007
We describe in this paper two methods for 3D shape indexing and retrieval that we apply on two data collections of the SHREC - SHape Retrieval Contest 2007: Watertight models and 3D CAD models.
Adamek, Tomasz, Napoléon, Thibault, O'Connor, Noel E., Schmitt, Francis   +3 more
core  

Hybrid CNN and Dictionary-Based Models for Scene Recognition and Domain Adaptation

, 2016
Convolutional neural network (CNN) has achieved state-of-the-art performance in many different visual tasks. Learned from a large-scale training dataset, CNN features are much more discriminative and accurate than the hand-crafted features. Moreover, CNN
Liu, Cheng-Lin, Xie, Guo-Sen, Yan, Shuicheng, Zhang, Xu-Yao   +3 more
core   +1 more source

Multi-Scale Fusion for Object Representation

Representing images or videos as object-level feature vectors, rather than pixel-level feature maps, facilitates advanced visual tasks. Object-Centric Learning (OCL) primarily achieves this by reconstructing the input under the guidance of Variational Autoencoder (VAE) intermediate representation to drive so-called slots to aggregate as much object ...
Pajarinen Joni, Kannala Juho, Zhao Rongzhen, Wang Vivienne   +3 more
openaire   +6 more sources

MSDSANet: Multimodal Emotion Recognition Based on Multi-Stream Network and Dual-Scale Attention Network Feature Representation

Sensors
Aiming at the shortcomings of EEG emotion recognition models in feature representation granularity and spatiotemporal dependence modeling, a multimodal emotion recognition model integrating multi-scale feature representation and attention mechanism is ...
Weitong Sun, Xingya Yan, Yuping Su, Gaihua Wang, Yumei Zhang   +4 more
doaj   +1 more source

Image fusion method based on simultaneous sparse representation with non‐subsampled contourlet transform

IET Computer Vision, 2019
The image fusion method based on sparse representation in the single‐scale image domain has produced better fusion results than the classic methods based on multi‐scale analysis nowadays.
Guiqing He, Siyuan Xing, Xingjian He, Jun Wang, Jianping Fan   +4 more
doaj   +1 more source

Underwater organisms detection algorithm based on multi‐scale perception and representation enhancement

IET Image Processing
To address issues such as object‐background confusion and difficulties in multi‐scale object feature extraction in underwater scenarios, this article proposes an underwater organisms detection algorithm based on multi‐scale perception and representation ...
Jiawei Xu, Fen Chen, Lian Huang, Tingna Liu, Zongju Peng   +4 more
doaj   +1 more source

Tailed U-Net: Multi-Scale Music Representation Learning

, 2022
Self-supervised learning has steadily been gaining traction in recent years. In music information retrieval (MIR), one promising recent application of self-supervised learning is the CLMR framework (contrastive learning of musical representations). CLMR has shown good performance, achieving results on par with state-of-the-art end-to-end classification
Vélez Vásquez, M.A., Burgoyne, J.A.
openaire   +3 more sources

Multi-Scale Representation of Ocean Flow Fields Based on Feature Analysis

ISPRS International Journal of Geo-Information, 2020
When it comes to feature retention in multi-scale representations of ocean flow fields, not all data points are equal. Therefore, this paper proposes a method of selecting data points based on their importance.
Bo Ai, Decheng Sun, Yanmei Liu, Chengming Li, Fanlin Yang, Yong Yin, Huibo Tian   +6 more
doaj   +1 more source

Multiscale Feature Extraction and Fusion of Image and Text in VQA

International Journal of Computational Intelligence Systems, 2023
The Visual Question Answering (VQA) system is the process of finding useful information from images related to the question to answer the question correctly.
Siyu Lu, Yueming Ding, Mingzhe Liu, Zhengtong Yin, Lirong Yin, Wenfeng Zheng   +5 more
doaj   +1 more source