Estimating Cement Compressive Strength from Microstructure Images Using Broad Learning System

Dang, Yonghao; Wang, Lin; Yin, Jianqin; Zhu, Xuehui; Feng, Zhiquan; Guo, Jifeng

doi:10.1109/smc.2018.00716

Cited by 10 publications

(3 citation statements)

References 26 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Liu and Zhou et al [22] employed a novel broad learning structure based on the K-means to cluster images in CIFAR-10 dataset. Dang and Wang et al [28] applied broad learning system to estimate cement compressive strength from microstructure images of cement. In [29], BLS was utilized for estimating sun visibility from given outdoor images, and then incremental broad learning system, the improvement of broad learning system proposed in [20], was applied to classify the sun visibility into more categories.…”

Section: B Applications Of Bls In Image Processingmentioning

confidence: 99%

DWnet: Deep-wide network for 3D action recognition

Dang

Yin

2020

Robotics and Autonomous Systems

Self Cite

View full text Add to dashboard Cite

We propose in this paper a deepwide network (DWnet) which combines the deep structure with the broad learning system (BLS) to recognize actions. Compared with the deep structure, the novel model saves lots of testing time and almost achieves realtime testing. Furthermore, the DWnet can capture better features than broad learning system can. In terms of methodology, we use pruned hierarchical co-occurrence network (PruHCN) to learn local and global spatialtemporal features. To obtain sufficient global information, BLS is used to expand features extracted by PruHCN. Experiments on two common skeletal datasets demonstrate the advantage of the proposed model on testing time and the effectiveness of the novel model to recognize the action.

show abstract

Section: B Applications Of Bls In Image Processingmentioning

confidence: 99%

DWnet: Deep-wide network for 3D action recognition

Dang

Yin

2020

Robotics and Autonomous Systems

Self Cite

View full text Add to dashboard Cite

show abstract

“…Deep convolutional neural network (DCNN) is a class of artificial neural network in deep learning, most adopted for 2D image analysis. The application of DCNN to microstructure-based property prediction − and phase segmentation , has demonstrated a huge potential in analyzing the highly disordered microstructure of cement paste. For example, the compressive strength and − creep modulus can be predicted using DCNN by analyzing the microstructure images of cementitious material.…”

Section: Introductionmentioning

confidence: 99%

“…The application of DCNN to microstructure-based property prediction − and phase segmentation , has demonstrated a huge potential in analyzing the highly disordered microstructure of cement paste. For example, the compressive strength and − creep modulus can be predicted using DCNN by analyzing the microstructure images of cementitious material. Besides, literature , has also suggested that different phases within the cement matrix, such as voids, sands, and fibers, can be segmented using DCNN.…”

Section: Introductionmentioning

confidence: 99%

Revealing Microstructural Modifications of Graphene Oxide-Modified Cement via Deep Learning and Nanoporosity Mapping: Implications for Structural Materials’ Performance

Wang

Chen

Duan

et al. 2022

ACS Appl. Nano Mater.

View full text Add to dashboard Cite

Locating, characterizing, and understanding the microstructural nanomodifications of random heterogeneous materials are critical to improve their performance, but current methods are limited by the material’s highly disordered microstructure. Here, we proposed a scheme to reveal the hidden microstructural modifications by coupling large-scale nanoporosity mappings with deep learning and demonstrated its effectiveness on nanomaterial (graphene oxide, GO)-modified cement. This deep learning-based approach showed superior abilities in distinguishing the nanomodified samples. The deep learning-approach shows a classification accuracy of 88.8% using nanoporosity mapping with micro-scale features. When nano-scale features were included, the classification accuracy further increased to 98%, indicating the GO modifications in the nano-scale. The microstructural changes in nano and micro scales were also located, providing the first direct evidence of a localized modification effect of GO where a small proportion (3.6–5.4%) of characteristic regions contribute up to 70% of total relevance. The effect of GO on spatial heterogeneity was also revealed by the changes in characteristic lengths. Besides, the reduced fraction of shared pore structural patterns from 44.6 to 40.9% under a high dosage of GO (0.02 wt %) also indicated the nanomodification effect of GO in improving structural topological efficiency. This study not only provides insights into the GO-modified cement for structural applications but also promotes the development of other random heterogeneous materials for a wide range of applications such as load-bearing, thermal and electrical insulation, catalyst support, and so on.

show abstract