Multimodal fusion for multimedia analysis: a survey

By Pradeep Kumar Atrey, M. Anwar Hossain, Abdulmotaleb El Saddik, Mohan Kankanhalli. Multimedia Systems, 2010

1,297 citations recorded by the source corpus.

Subfields: Biomedical Engineering, Media Technology, Signal Processing

View Paper: https://www.semanticscholar.org/paper/e9090508b9073763c6693983577137db2a41a25b

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Biomedical Engineering

In Media Technology

In Signal Processing

Citations in the corpus

Multimodal Machine Learning: A Survey and Taxonomy (2017, background): “A second important category of multimodal applications comes from the field of multimedia content indexing and retrieval [11], [188].”
Trusted Multi-View Classification With Dynamic Evidential Fusion (2022, background): “DS evidence theory has also been extensively studied in multiview learning, which mainly focuses on improving the Dempster’s combination rule [17], [76], [77], [78], [79], [80].”
Multimodal Classification Fusion in Real-World Scenarios (2017, methodology): “The text from the dataset was represented as a BoW vector of size 804, that is the full size of the dictionary.”
Exploiting Feature and Class Relationships in Video Categorization with Regularized Deep Neural Networks (2015, background): “Simple fusion strategies like the early or the late fusion usually have limited performance improvement since the intrinsic relations among multiple feature representations are often overlooked [46].”
Fusing Multiple Features for Depth-Based Action Recognition (2015, neutral): “Replacing it with summation can result in much better results: 18:10 W , b , 2 The SVM-based fusion is based on a two-layer structure, where the input to the higher-layer SVM the conﬁdence scores given by individual lower-layer SVM classiﬁers [Atrey et al. 2010].”
Deep Multimodal Fusion by Channel Exchanging (2020, background): “Another categorization of multimodal fusion can be specified as early, middle, and late fusion, depending on when to fuse, which have been discussed in earlier works [2, 7, 17, 41] and also in the deep learning literature [4, 26, 27, 45].”
Conversational Memory Network for Emotion Recognition in Dyadic Dialogue Videos (2018, background): “Literature consists of numerous fusion techniques for multimodal data (Atrey et al., 2010; Zadeh et al., 2017; Poria et al., 2017c).”
Joint Feature Selection and Subspace Learning for Cross-Modal Retrieval (2016, background): “Although much attention has been paid to multimodal data analysis [1], [2], [3], [4], the common strategy is to integrate multiple modalities to improve”
Gated Multimodal Units for Information Fusion (2017, background): “Different reviews (Atrey et al., 2010; Bhatt & Kankanhalli, 2011; Li Deng, 2014; Deng, 2014) have summarized strategies that addressed multimodal analysis.”
A Review of Human Activity Recognition Methods (2015, background): “In the literature, there are two main fusion strategies that can be used to tackle this problem (Atrey et al., 2010; Shivappa et al., 2010). Early fusion, or fusion at the feature level, combines features of different modalities, usually by reducing the dimensionality in eachmodality and creating a ”
Deep Multimodal Learning: A Survey on Recent Advances and Trends (2017, background): “Techniques for multimodal data fusion, which cover different application domains, have long been investigated by the research community [3], [4].”
Content-based visual search learned from social media (2012, methodology): “C o m p ared w th th e b aelin e [5 2 ], w e o b tain a retive im p ro em en t in term s o f P @ 2 0 b y 1 5 .”
A Survey of Multi-View Representation Learning (2016, neutral): “Recently, several related surveys [4], [24], [25], [26], [27] of”