Wyniki wyszukiwania - BazTech

1

AUV Fault Diagnosis Based on Multidimensional Temporal Classification Transformer

Chen Weifeng, Xu Xiaoyan, Chen Mingzhi, Tarasiuk Tomasz

Polish Maritime Research

|

2025

|

nr 4

97--106

EN

Autonomous underwater vehicles (AUVs) play a critical role in marine resource exploration, maritime patrol, and rescue operations, requiring reliable fault diagnosis for robust operation. This paper proposes a novel AUV fault diagnosis method based on the multidimensional temporal classification transformer (MTC-Transformer) deep learning algorithm. The MTC-Transformer enhances traditional transformer architectures by optimising them for multidimensional time-series data processing and fault classification. It features adaptive automatic feature extraction, superior multi-modal data handling, and improved scalability. Key innovations include gated fusion for combining features from outlier-processed data and kernel principal component analysis-reduced time-series data, a dedicated fault feature amplification mechanism, and a multi-layer perceptron head for classification. Trained and validated on the ‘Haizhe’ small quadrotor AUV fault dataset using double-layer randomised sampling to ensure generalisation, the model achieves exceptional accuracies of 99.51% (training) and 99.44% (validation). Comparative experiments demonstrate its superiority over WDCNN, LSTM-1DCNN, and other benchmarks in handling variable-length sequences and capturing long-range dependencies, confirming its high accuracy, robustness, and practical efficacy for AUV fault diagnosis.

2

FRE-Net: Full-region enhanced network for nuclei segmentation in histopathology images

Huang Xuping, Chen Junxi, Chen Mingzhi, Wan Yaping, Chen Lingna

Biocybernetics and Biomedical Engineering

|

2023

|

Vol. 43, no. 1

386--401

EN

Accurate nuclei segmentation is a critical step for physicians to achieve essential information about a patient’s disease through digital pathology images, enabling an effective diagnosis and evaluation of subsequent treatments. Since pathology images contain many nuclei, manual segmentation is time-consuming and error-prone. Therefore, developing a precise and automatic method for nuclei segmentation is urgent. This paper proposes a novel multi-task segmentation network that incorporates background and contour segmentation into the nuclei segmentation method and produces more accurate segmentation results. The convolution and attention modules are merged with the model to increase its global focus and enhance good segmentation results indirectly. We propose a reverse feature enhance module for contour extraction that facilitates feature integration between auxiliary tasks. The multi-feature fusion module is embedded in the final decoding branch to use different levels of features from auxiliary segmentation branches with varying concerns. We evaluate the proposed method on four challenging nuclei segmentation datasets. The proposed method achieves excellent performance on all four datasets. We found that the Dice coefficient reached 0.8563±0.0323, 0.8183±0.0383, 0.9222±0.0216, and 0.9220±0.0602 on the TNBC, MoNuSeg, KMC, and Glas. Our method produces better boundary accuracy and less sticking than other end-to-end segmentation methods. The results show that our method can perform better than other proposed state-of-the-art methods.

3

A dual-stage transformer and MLP-based network for breast ultrasound image segmentation

Lin Guidi, Chen Mingzhi, Tan Minsheng, Chen Lingna, Chen Junxi

Biocybernetics and Biomedical Engineering

|

2023

|

Vol. 43, no. 4

656--671

EN

Automatic segmentation of breast lesions from ultrasound images plays an important role in computer-aided breast cancer diagnosis. Many deep learning methods based on convolutional neural networks (CNNs) have been proposed for breast ultrasound image segmentation. However, breast ultrasound image segmentation is still challenging due to ambiguous lesion boundaries. We propose a novel dual-stage framework based on Transformer and Multi-layer perceptron (MLP) for the segmentation of breast lesions. We combine the Swin Transformer block with an efficient pyramid squeezed attention block in a parallel design and introduce bi-directional interactions across branches, which can efficiently extract multi-scale long-range dependencies to improve the segmentation performance and robustness of the model. Furthermore, we introduce tokenized MLP block in the MLP stage to extract global contextual information while retaining fine-grained information to segment more complex breast lesions. We have conducted extensive experiments with state-of-the-art methods on three breast ultrasound datasets, including BUSI, BUL, and MT_BUS datasets. The dice coefficient reached 0.8127 ± 0.2178, and the intersection over union reached 0.7269 ± 0.2370 on benign lesions when the Hausdorff distance was maintained at 3.75 ± 1.83. The dice coefficient of malignant lesions is improved by 3.09% for BUSI dataset. The segmentation results on the BUL and MT_BUS datasets also show that our proposed model achieves better segmentation results than other methods. Moreover, the external experiments indicate that the proposed model provides better generalization capability for breast lesion segmentation. The dual-stage scheme and the proposed Transformer module achieve the fine-grained local information and long-range dependencies to relieve the burden of radiologists.

4

Monitoring the performance of a ship’s main engine based on big data technology

Liang Meng, Chen Mingzhi

Polish Maritime Research

|

2022

|

nr 3

128--140

EN

Under the recent background of ‘Green Shipping’ and rising fuel prices, it is very important to reduce the fuel consumption rate of ships, which is directly affected by the performance of the main engine. A reasonable maintenance schedule can optimise the performance of the main engine. However, a traditional maintenance schedule is based on the navigation distance and time, ignoring many other factors, such as a harsh working environments and frequently changing operating conditions, which will lead to faster performance degradation. In this study, a real-time evaluation method combing big data of ship energy efficiency with physics-based analysis is proposed to judge the degradation of main engine performance and assist in determining the maintenance schedule. Firstly, based on the developed ship energy efficiency big data platform, the distribution statistics and comparison of different operating states are carried out. Gaussian mixture model (GMM) and Density-Based Spatial Clustering of Applications with Noise (DBSCAN) are used to cluster the data and the high-density data areas are obtained as the analysis points. Then, the data of the analysis points are polynomial fitted, by the least square method, to obtain the propulsion characteristics curves, load characteristic curves, and speed characteristic curves, which can be used to observe the performance degradation of the main engine. The results show that this method can effectively monitor the degradation degree of the main engine performance, and is of great significance to fuel efficiency improvements and greenhouse gas (GHG) emissions reduction.