Deep Learning Based Dimple Segmentation for Quantitative Fractography

Sinha, Ashish; Suresh, K.S.

doi:10.1007/978-3-030-68799-1_34

Cited by 9 publications

(4 citation statements)

References 22 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The DSC evaluation metric improves from 85.37% to 94.65%, and HD95 decreases from 85.76 to 22.99. Sinha's work realized the segmentation of the obvious deep regions in dimples [33]. Different from Sinha's model, seen in Figure 5c,f, Swin-UNet++ could detect dimples accurately and also segment the entire morphology of dimples precisely on the fracture surface.…”

Section: Training and Segmentation Resultsmentioning

confidence: 91%

“…This model can well segment the clearly visible deep dimples. However, this model is inappropriate to segment the overall dimple morphology of fracture [33].…”

Section: Introductionmentioning

confidence: 99%

“…The aforementioned limitations [11,32,33] are mainly caused due to the inability of deep learning methods to model the explicit long-range relationship and the intrinsic locality of the convolution operations in CNN-based models [34]. It is notable that, usually, the distribution of the dimples is not concentrated on the fractured surface.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Swin–UNet++: A Nested Swin Transformer Architecture for Location Identification and Morphology Segmentation of Dimples on 2.25Cr1Mo0.25V Fractured Surface

et al. 2021

View full text Add to dashboard Cite

The precise identification of micro-features on 2.25Cr1Mo0.25V steel is of great significance for understanding the mechanism of hydrogen embrittlement (HE) and evaluating the alloy’s properties of HE resistance. Presently, the convolution neural network (CNN) of deep learning is widely applied in the micro-features identification of alloy. However, with the development of the transformer in image recognition, the transformer-based neural network performs better on the learning of global and long-range semantic information than CNN and achieves higher prediction accuracy. In this work, a new transformer-based neural network model Swin‒UNet++ was proposed. Specifically, the architecture of the decoder was redesigned to more precisely detect and identify the micro-feature with complex morphology (i.e., dimples) of 2.25Cr1Mo0.25V steel fracture surface. Swin‒UNet++ and other segmentation models performed state-of-the-art (SOTA) were compared on the dimple dataset constructed in this work, which consists of 830 dimple scanning electron microscopy (SEM) images on 2.25Cr1Mo0.25V steel fracture surface. The segmentation results show Swin‒UNet++ not only realizes the accurate identification of dimples but displays a much higher prediction accuracy and stronger robustness than Swin‒Unet and UNet. Moreover, efforts from this work will also provide an important reference value to the identification of other micro-features with complex morphologies.

show abstract

Section: Training and Segmentation Resultsmentioning

confidence: 91%

“…This model can well segment the clearly visible deep dimples. However, this model is inappropriate to segment the overall dimple morphology of fracture [33].…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Swin–UNet++: A Nested Swin Transformer Architecture for Location Identification and Morphology Segmentation of Dimples on 2.25Cr1Mo0.25V Fractured Surface

et al. 2021

View full text Add to dashboard Cite

show abstract

“…The voids are closer together, and the total gap volume is larger when the external load acting on the specimen is more. 29 It should be noted that void nucleation, coalescence of grains and honeycomb with river pattern are all known to initiate on the microcracks. Morphologies such as voids, dimples, honeycombs and riverine structures are observed (Figure 6).…”

Section: Discussionmentioning

confidence: 99%

Additive manufacturing and characterization of titanium wall used in nuclear application

Veeman¹,

Subramaniyan²,

Kumar

et al. 2023

Proceedings of the Institution of Mechanical Engineers, Part L:

View full text Add to dashboard Cite

Titanium and its alloy are attractive materials used worldwide for various applications, especially for manufacturing fusion reactors and pressure vessels. This study aims to provide knowledge on the microstructural and mechanical characteristics of titanium grade 9 wall fabricated using wire feeding additive manufacturing (WFAM) with gas tungsten arc welding (GTAW)-aided technique. An inert environment is provided during the process to manufacture a defect-free wall. Due to varying thermal cycles, there is a variation in the microstructure of the WFAM processed wall. The change in the texture of the WFAM processed wall is analysed through electron backscatter diffraction spectroscopy (inverse pole figures and pole figures), which influences the material properties. The WFAM processed wall exhibits superior mechanical characteristics compared to wrought alloys. Numerous morphologies in the fractured surface infer the occurrence of ductile fracture. A comparative study has been conducted between the outcomes of finite element simulation and experimental results. The results of finite element simulation and experimentation were in good agreement. The compilation of results gathered from mechanical and microstructural characterization shows that gas tungsten arc welding-aided-WFAM titanium grade 9 wall is suitable for fabricating fusion reactors and pressure vessels. Hence, first-hand knowledge is gained on the fabrication of fusion reactors and pressure vessels in nuclear application using the tungsten arc welding-aided-WFAM titanium grade 9 wall.

show abstract

Methods for evaluating fracture patterns of polycrystalline materials based on the parameter analysis of ductile separation dimples: A review

Maruschak,

Konovalenko,

Sorochak

2023

Engineering Failure Analysis

View full text Add to dashboard Cite

Deep Learning Based Dimple Segmentation for Quantitative Fractography

Cited by 9 publications

References 22 publications

Swin–UNet++: A Nested Swin Transformer Architecture for Location Identification and Morphology Segmentation of Dimples on 2.25Cr1Mo0.25V Fractured Surface

Swin–UNet++: A Nested Swin Transformer Architecture for Location Identification and Morphology Segmentation of Dimples on 2.25Cr1Mo0.25V Fractured Surface

Additive manufacturing and characterization of titanium wall used in nuclear application

Methods for evaluating fracture patterns of polycrystalline materials based on the parameter analysis of ductile separation dimples: A review

Contact Info

Product

Resources

About