Towards autonomous and optimal excavation of shield machine: a deep reinforcement learning-based approach

Zhang, Yakun; Gong, Gwo‐Ching; Yang, Huayong; Chen, Yuxi; Chen, Genglin

doi:10.1631/jzus.a2100325

Cited by 13 publications

(3 citation statements)

References 49 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Based on region segmentation, in 2018, Zhang et al. ( 16 ) used cam-based localization of the lesion region, but only general localization of the lesion was carried out without a specific contour. In 2020, Xue et al.…”

Section: Related Workmentioning

confidence: 99%

Segmentation of the cervical lesion region in colposcopic images based on deep learning

Fan

et al. 2022

Front. Oncol.

View full text Add to dashboard Cite

BackgroundColposcopy is an important method in the diagnosis of cervical lesions. However, experienced colposcopists are lacking at present, and the training cycle is long. Therefore, the artificial intelligence-based colposcopy-assisted examination has great prospects. In this paper, a cervical lesion segmentation model (CLS-Model) was proposed for cervical lesion region segmentation from colposcopic post-acetic-acid images and accurate segmentation results could provide a good foundation for further research on the classification of the lesion and the selection of biopsy site.MethodsFirst, the improved Faster Region-convolutional neural network (R-CNN) was used to obtain the cervical region without interference from other tissues or instruments. Afterward, a deep convolutional neural network (CLS-Net) was proposed, which used EfficientNet-B3 to extract the features of the cervical region and used the redesigned atrous spatial pyramid pooling (ASPP) module according to the size of the lesion region and the feature map after subsampling to capture multiscale features. We also used cross-layer feature fusion to achieve fine segmentation of the lesion region. Finally, the segmentation result was mapped to the original image.ResultsExperiments showed that on 5455 LSIL+ (including cervical intraepithelial neoplasia and cervical cancer) colposcopic post-acetic-acid images, the accuracy, specificity, sensitivity, and dice coefficient of the proposed model were 93.04%, 96.00%, 74.78%, and 73.71%, respectively, which were all higher than those of the mainstream segmentation model.ConclusionThe CLS-Model proposed in this paper has good performance in the segmentation of cervical lesions in colposcopic post-acetic-acid images and can better assist colposcopists in improving the diagnostic level.

show abstract

Section: Related Workmentioning

confidence: 99%

Segmentation of the cervical lesion region in colposcopic images based on deep learning

Fan

et al. 2022

Front. Oncol.

View full text Add to dashboard Cite

show abstract

“…With more challenges, judges also face more dilemmas in the process of adapting to the new changes brought about by informatization: First, the cases are becoming more and more complex, and the results of the judgment are difficult to grasp and unify; second, there are a lot of relevant materials in the case, which makes it difficult to sort out and check the facts of the case [3]. Legal judgment prediction technology alleviates the above problems to a certain extent, and is a key technology of the intelligent legal assistant system [4].…”

Section: Introductionmentioning

confidence: 99%

Problem Analysis and Application Research in Legal Judgment Prediction Based on Data Mining

2024

jes

View full text Add to dashboard Cite

The judgments of some hot cases that exceeded public expectations seem to confirm this. Although the legal certainty is questioned, the judgment remains predictable. The predictability of judgments is a way to enhance judicial authority and maintain judicial credibility. It is also a way to achieve overlapping consensus between the judiciary and the public, provide stable value guidance and behavioral expectations for the public, and promote the generation and development of public rational trust. To enhance public legal trust by improving the predictability of judgments, it is necessary to increase the burden of reasoning for judicial judgments, avoid randomness and contingency, and ensure the adequate provision and substantial disclosure of previous judgment information, so that judgment prediction in the era of big data can truly become possible. The research is to use the optimized particle swarm algorithm as the underlying model to carry out joint modeling and prediction research on the analysis and application of problems in the prediction of legal judgments. According to experimental calculations, the optimized particle swarm algorithm can significantly improve the accuracy and universality of legal judgment prediction. After optimization, the convergence speed is increased by about 5%, and the value of the acceleration factor is more significant.

show abstract

“…Secondly, the high integration level of TBM processes necessitates advanced technical skills and experience from operators, where improper handling could not only impair boring efficiency but also risk mechanical failures or accidents. Consequently, many researchers are now turning their attention to the development of automated and intelligent technologies for TBM operations [1][2][3][4][5].…”

Section: Introductionmentioning

confidence: 99%

Semi-Symmetrical, Fully Convolutional Masked Autoencoder for TBM Muck Image Segmentation

Lei,

Tan,

Wang

et al. 2024

Symmetry

View full text Add to dashboard Cite

Deep neural networks are effectively utilized for the instance segmentation of muck images from tunnel boring machines (TBMs), providing real-time insights into the surrounding rock condition. However, the high cost of obtaining quality labeled data limits the widespread application of this method. Addressing this challenge, this study presents a semi-symmetrical, fully convolutional masked autoencoder designed for self-supervised pre-training on extensive unlabeled muck image datasets. The model features a four-tier sparse encoder for down-sampling and a two-tier sparse decoder for up-sampling, connected via a conventional convolutional neck, forming a semi-symmetrical structure. This design enhances the model’s ability to capture essential low-level features, including geometric shapes and object boundaries. Additionally, to circumvent the trivial solutions in pixel regression that the original masked autoencoder faced, Histogram of Oriented Gradients (HOG) descriptors and Laplacian features have been integrated as novel self-supervision targets. Testing shows that the proposed model can effectively discern essential features of muck images in self-supervised training. When applied to subsequent end-to-end training tasks, it enhances the model’s performance, increasing the prediction accuracy of Intersection over Union (IoU) for muck boundaries and regions by 5.9% and 2.4%, respectively, outperforming the enhancements made by the original masked autoencoder.

show abstract

Towards autonomous and optimal excavation of shield machine: a deep reinforcement learning-based approach

Cited by 13 publications

References 49 publications

Segmentation of the cervical lesion region in colposcopic images based on deep learning

Segmentation of the cervical lesion region in colposcopic images based on deep learning

Problem Analysis and Application Research in Legal Judgment Prediction Based on Data Mining

Semi-Symmetrical, Fully Convolutional Masked Autoencoder for TBM Muck Image Segmentation

Contact Info

Product

Resources

About