Enhancing Anomaly Detection for Cultural Heritage via Long Short-Term Memory with Attention Mechanism

被引:2
作者
Wu, Yuhan [1 ]
Dong, Yabo [1 ]
Shan, Zeyang [1 ]
Meng, Xiyu [1 ]
He, Yang [1 ]
Jia, Ping [2 ]
Lu, Dongming [1 ]
机构
[1] Zhejiang Univ, Coll Comp Sci & Technol, Hangzhou 310027, Peoples R China
[2] China Mobile Grp Zhejiang Co Ltd, Hangzhou 311100, Peoples R China
关键词
cultural heritage; anomaly warning; LSTM; attention mechanism; timber structure; DETERIORATION; PREDICTION; RISK;
D O I
10.3390/electronics13071254
中图分类号
TP [自动化技术、计算机技术];
学科分类号
0812 ;
摘要
Cultural heritages are invaluable and non-renewable resources. Existing warning mechanisms usually rely on degradation analysis to assess environmental risk factors. However, they have limitations such as complex research, poor generalization, and inadequate warnings. To address these issues, we propose a hybrid model that combines the long short-term memory network (LSTM) and attention mechanisms with environmental factors to detect anomalies in cultural heritage. The attention mechanism extracts temporal dependencies, while LSTM captures global long-term patterns, enhancing the reconstruction capability. Additionally, we introduce the seasonal, trend, shapelet, and mixed anomalies to enrich the positive and negative samples and propose a novel threshold extraction method to reduce the reliance on prior knowledge. Extensive experiments demonstrate that LSTM-Attention outperforms previous methods, with a root mean square error (RMSE) of 34.328, mean absolute error (MAE) of 27.060, and the largest area under the receiver operating characteristic curve (AUC) value, highlighting the superiority and effectiveness of our method. The ablation study proves the contribution of the attention mechanism, and the reconstruction step can be flexibly adjusted according to practical needs. Finally, the algorithm has been successfully implemented in engineering and employed in practice, providing valuable guidelines for the preservation of ancient buildings worldwide.
引用
收藏
页数:21
相关论文
共 53 条
[1]  
Akcaozoglu S, 2020, FRESEN ENVIRON BULL, V29, P2338
[2]  
Betti M, 2016, EARTHQUAKES THEIR IM, P377, DOI [10.1007/978-3-319-21753-614, DOI 10.1007/978-3-319-21753-6_14]
[3]   LSTM model for visual speech recognition through facial expressions [J].
Bhaskar, Shabina ;
Thasleema, T. M. .
MULTIMEDIA TOOLS AND APPLICATIONS, 2023, 82 (04) :5455-5472
[4]   Sentiment analysis for user reviews using Bi-LSTM self-attention based CNN model [J].
Bhuvaneshwari, P. ;
Rao, A. Nagaraja ;
Robinson, Y. Harold ;
Thippeswamy, M. N. .
MULTIMEDIA TOOLS AND APPLICATIONS, 2022, 81 (09) :12405-12419
[5]  
Box G.E.P., 1970, Time series analysis: forecasting and control, DOI 10.1080/01621459.1970.10481180
[6]   Estimating bank default with generalised extreme value regression models [J].
Calabrese, Raffaella ;
Giudici, Paolo .
JOURNAL OF THE OPERATIONAL RESEARCH SOCIETY, 2015, 66 (11) :1783-1792
[7]   Word Acquisition in Neural Language Models [J].
Chang, Tyler A. ;
Bergen, Benjamin K. .
TRANSACTIONS OF THE ASSOCIATION FOR COMPUTATIONAL LINGUISTICS, 2022, 10 :1-16
[8]   AutoFormer: Searching Transformers for Visual Recognition [J].
Chen, Minghao ;
Peng, Houwen ;
Fu, Jianlong ;
Ling, Haibin .
2021 IEEE/CVF INTERNATIONAL CONFERENCE ON COMPUTER VISION (ICCV 2021), 2021, :12250-12260
[9]   Learning Graph Structures With Transformer for Multivariate Time-Series Anomaly Detection in IoT [J].
Chen, Zekai ;
Chen, Dingshuo ;
Zhang, Xiao ;
Yuan, Zixuan ;
Cheng, Xiuzhen .
IEEE INTERNET OF THINGS JOURNAL, 2022, 9 (12) :9179-9189
[10]  
Chikobvu Delson, 2015, S. Afr. j. sci., V111, P01, DOI 10.17159/SAJS.2015/20140271