Real-Time Terahertz Characterization of Minor Defects by the YOLOX-MSA Network

Wang, Xingyu; Zhang, Zhen; Xu, Yafei; Zhang, Liuyang; Yan, Ruqiang; Chen, Xuefeng

doi:10.1109/tim.2022.3201945

Cited by 12 publications

(2 citation statements)

References 19 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The performance of three neural network models have been compared and the 1D CNN outperformed LSTM RNN and bidirectional LSTM RNN models based on recall and F1 score (harmonic mean of precision and recall scores i.e., 𝐹1 = 2( × )) metrics. The YOLO-MSA network has been investigated for detection of minor defects with high accuracy at real time speed for industrial online detection systems [108] and the YOLO V4 used for printed circuit board (PCB) defects detection in THz nondestructive testing application [109]. The recognition of tissue burns has been investigated using CNN which proved to be more robust than existing algorithms in [110].…”

Section: B Classification Detection and Identificationmentioning

confidence: 99%

Terahertz Data Extraction and Analysis Based on Deep Learning Techniques for Emerging Applications

Gezimati,

Singh

2024

IEEE Access

View full text Add to dashboard Cite

Following the recent progress in the development of Terahertz (THz) generation and detection, THz technology is being widely used to characterize test sample properties in various applications including nondestructive testing, security inspection and medical applications. In this paper, we have presented a broad review of the recent usage of artificial intelligence (AI) particularly, deep learning techniques in various THz sensing, imaging, and spectroscopic applications with emphasis on their implementation for medical imaging of cancerous cells. Initially, the fundamentals principles and techniques for THz generation and detection, imaging and spectroscopy are introduced. Subsequently, a brief overview of AI -machine learning and deep learning techniques is summarized, and their performance is compared. Further, the usage of deep learning algorithms in various THz applications is reported, with focus on metamaterials design and classification, detection, reconstruction, segmentation, parameter extraction and denoising tasks. Moreover, we also report the metrics used to evaluate the performance of deep learning models and finally, the existing research challenges in the application of deep learning in THz cancer imaging applications are identified and possible solutions are suggested through emerging trends. With the continuous increase of acquired THz data -sensing, spectral and imaging, artificial intelligence has emerged as a dominant paradigm for embedded data extraction, understanding, perception, decision making and analysis. Towards this end, the integration of state-of-the-art machine learning techniques such as deep learning with THz applications enable detailed computational and theoretical analysis for better validation and verification than modelling techniques that precede the era of machine learning. The study will facilitate the large-scale clinical applications of deep learning enabled THz imaging systems for the development of smart and connected next generation healthcare systems as well as provide a roadmap for future research direction.

show abstract

Section: B Classification Detection and Identificationmentioning

confidence: 99%

Terahertz Data Extraction and Analysis Based on Deep Learning Techniques for Emerging Applications

Gezimati,

Singh

2024

IEEE Access

View full text Add to dashboard Cite

show abstract

“…MobileNet [35] was used as the backbone, and they introduced attention mechanism modules such as local self-attention architecture (HaloNet) [36] and squeezeand-excitation network (SENet) [37] in order to gain a large receptive field. Wang et al [38] via the way to replace Swin Transformer [39] as the backbone of YOLOX, also modified the detection heads and loss function in the Terahertz detection. Gao et al [40] verified the effectiveness of defect detection in the Swin Transformer and finally proposed Cas-VSwin Transformer.…”

Section: Application Development In Industrymentioning

confidence: 99%

AMFF-YOLOX: Towards an Attention Mechanism and Multiple Feature Fusion Based on YOLOX for Industrial Defect Detection

et al. 2023

View full text Add to dashboard Cite

Industrial defect detection has great significance in product quality improvement, and deep learning methods are now the dominant approach. However, the volume of industrial products is enormous and mainstream detectors are unable to maintain a high accuracy rate during rapid detection. To address the above issues, this paper proposes AMFF-YOLOX, an improved industrial defect detector based on YOLOX. The proposed method can reduce the activation function and normalization operation of the bottleneck in the backbone network, and add an attention mechanism and adaptive spatial feature fusion within the feature extraction network to enable the network to better focus on the object. Ultimately, the accuracy of the prediction is enhanced without excessive loss of speed in network prediction, with competitive performance compared to mainstream detectors. Experiments show that the proposed method in this paper achieves 61.06% (85.00%) mAP@0.5:0.95 (mAP@0.5) in the NRSD-MN dataset, 51.58% (91.09%) is achieved in the PCB dataset, and 49.08% (80.48%) is achieved in the NEU-DET dataset. A large number of comparison and ablation experiments validate the effectiveness and competitiveness of the model in industrial defect detection scenarios.

show abstract

THz Signal Identification for Intelligent Characterization Under High-Resolution Mode based on RFECNet

Wang,

Xu,

Wang

et al. 2023

2023 International Conference on Sensing, Measurement &Amp;amp; Data Analytics in the Era of Artificial Intelligence (ICSMD)

View full text Add to dashboard Cite

Real-Time Terahertz Characterization of Minor Defects by the YOLOX-MSA Network

Cited by 12 publications

References 19 publications

Terahertz Data Extraction and Analysis Based on Deep Learning Techniques for Emerging Applications

Terahertz Data Extraction and Analysis Based on Deep Learning Techniques for Emerging Applications

AMFF-YOLOX: Towards an Attention Mechanism and Multiple Feature Fusion Based on YOLOX for Industrial Defect Detection

THz Signal Identification for Intelligent Characterization Under High-Resolution Mode based on RFECNet

Contact Info

Product

Resources

About