Multimodal sentiment analysis has a wide range of applications due to its information complementarity in multimodal interactions. Previous works focus more on investigating efficient joint representations, but they rarely consider the insufficient unimodal features extraction and data redundancy of multimodal fusion. In this paper, a Video-based Cross-modal Auxiliary Network (VCAN) is proposed, which is comprised of an audio features map module and a cross-modal selection module. The first module is designed to substantially increase feature diversity in audio feature extraction, aiming to improve classification accuracy by providing more comprehensive acoustic representations. To empower the model to handle redundant visual features, the second module is addressed to efficiently filter the redundant visual frames during integrating audiovisual data. Moreover, a classifier group consisting of several image classification networks is introduced to predict sentiment polarities and emotion categories. Extensive experimental results on RAVDESS, CMU-MOSI, and CMU-MOSEI benchmarks indicate that VCAN is significantly superior to the state-of-the-art methods for improving the classification accuracy of multimodal sentiment analysis.
The transformation law of autogenous volume deformation of MgO-admixed concrete with specimen size was revealed through continuous observation on standard-size (Φ200 mm × 500 mm), medium-size (Φ250 mm × 500 mm), and large-size specimens (Φ250 mm × 600 mm) over 6 years. Besides, the pore parameters of concrete core samples obtained from autogenous volume deformation specimens in the 1st, 3rd, and 6th years of age were investigated. The results show that the autogenous volume deformation of MgO-admixed concrete increases with the increase in the MgO content or age. The expansion rate of the concrete specimen decreases after the age of 360 days, and the autogenous volume expansion deformation of the specimen tends to be stable after about 2 years. When the size of the specimen changes, the autogenous volume deformation of MgO-admixed concrete decreases with the increase of specimen size. During the age of 2–6 years, the expansion of medium- and large-size specimens is reduced by 6–10 and 15–20%, respectively, compared with the standard-size specimens under the same MgO content. With the condition of an appropriate MgO content, regardless of the size of the specimen, the pore structure of the concrete becomes better and better with the growth of age, the concrete becomes denser and denser, and the expansion caused by MgO hydration will not cause damage to concrete structures.
In order to accelerate the application of magnesium oxide (MgO) expanding admixture in roller compacted concrete (RCC), aiming at common three-grading RCC used in water conservancy and hydropower engineering, the autoclave expansion of mortar specimens and the compressive strength of the autoclaved mortar specimens were studied by means of autoclave test and compressive strength test. Results showed that the balanced MgO content in RCC, which cannot be determined by the MgO content corresponding to 0.5% autoclave expansion rate of mortar specimens or the inflection point of the curve of autoclave expansion changing with MgO content like normal concrete, should be determined by the MgO content corresponding to the inflection point of the curve that compressive strength of autoclaved mortar specimens changes with MgO content to ensure long-term safe operation of RCC dams.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.