2024
DOI: 10.1080/17452759.2024.2325572
|View full text |Cite
|
Sign up to set email alerts
|

AM-SegNet for additive manufacturing in situ X-ray image segmentation and feature quantification

Wei Li,
Rubén Lambert-Garcia,
Anna C. M. Getley
et al.
Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
2
1
1
1

Citation Types

0
7
0

Year Published

2024
2024
2024
2024

Publication Types

Select...
9

Relationship

0
9

Authors

Journals

citations
Cited by 17 publications
(7 citation statements)
references
References 48 publications
0
7
0
Order By: Relevance
“…These innovations could potentially enhance the performance of EndoAdd, particularly in challenging endoscopic environments where real-time processing and accurate detection are paramount. Moreover, the adoption of transformer-based models like DETR may offer new possibilities for handling the sequential nature of endoscopic video data, leveraging self-attention mechanisms for improved feature recognition and localization [18].…”
Section: Discussionmentioning
confidence: 99%
See 1 more Smart Citation
“…These innovations could potentially enhance the performance of EndoAdd, particularly in challenging endoscopic environments where real-time processing and accurate detection are paramount. Moreover, the adoption of transformer-based models like DETR may offer new possibilities for handling the sequential nature of endoscopic video data, leveraging self-attention mechanisms for improved feature recognition and localization [18].…”
Section: Discussionmentioning
confidence: 99%
“…Current AI models, such as convolutional neural networks (CNNs) integrated with long short-term memory (LSTM) networks and lightweight neural networks, have shown promising results in identifying operative phases in endoscopic procedures with high accuracy [18]. Yamazaki et al developed a system based on the YOLOv3 platform to detect and classify surgical instruments in laparoscopic gastrectomy videos.…”
Section: Introductionmentioning
confidence: 99%
“…In the field of AM, researchers continually develop innovative tools to ensure the quality of manufactured components. Li et al [69] introduced AM-SegNet, a deep learning model designed to rapidly and accurately segment high-resolution synchrotron X-ray images in metal AM. This model achieved an impressive 96% accuracy, with processing times below four milliseconds per frame, allowing for efficient identification and analysis of critical features like keyholes and pores, thereby advancing our understanding of AM processes.…”
Section: Importance Of Defect Analysis For Quality Controlmentioning
confidence: 99%
“…The U-Net architecture can take a pre-trained encoder and use it during training. U-Net architecture is widely used for semantic segmentation tasks, since the architecture can be modified and improved through different operations quite easily, which can lead to new network architectures based on U-Net such as those seen in [35,36]. For our research, the EfficientNet-b3 [37] encoder was selected and used for the multi-class and binary segmentation networks in this case.…”
Section: U-net Deep Convolutional Networkmentioning
confidence: 99%