3D ARCNN: An Asymmetric Residual CNN for False Positive Reduction in Pulmonary Nodule

Liu, Bowen; Song, Hong; Li, Qiang; Lin, Yucong; Weng, Xutao; Su, Zhaoyi; Yang, Jian

doi:10.1109/tnb.2023.3278706

Cited by 4 publications

(6 citation statements)

References 35 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Currently, studies (Gugulothu and Balaji, 2023;Halder and Dey, 2023;Jeyaraj and Nadar, 2023;Liu et al, 2023;Omar et al, 2023;Zhu et al, 2023) have made some progress in the task of lung tumor pathology classification. These methods include two main technical lines: lung cancer classification methods based on CT images (Gugulothu and Balaji, 2023;Liu et al, 2023;Zhu et al, 2023) and deep learning classification algorithms based on histopathological images (Halder and Dey, 2023;Jeyaraj and Nadar, 2023;Omar et al, 2023).…”

Section: Related Workmentioning

confidence: 99%

“…There are two main technical routes based on existing deep-learning techniques for lung cancer diagnosis and classification: 1. Lung cancer classification methods based on low-dose computed tomography (CT) images ( Anderson and Davis, 2018 ), which are mainly studied based on public datasets such as LIDC-IDRI and LUNA16 ( Gugulothu and Balaji, 2023 ; Liu et al, 2023 ; Zhu et al, 2023 ). 2.…”

Section: Introductionmentioning

confidence: 99%

“…Despite some progress made by works ( Gugulothu and Balaji, 2023 ; Halder and Dey, 2023 ; Jeyaraj and Nadar, 2023 ; Liu et al, 2023 ; Omar et al, 2023 ; Zhu et al, 2023 ) in the task of diagnostic classification of lung cancer, the pathological classification of lung tumors currently faces the following challenges: • The first is either only preliminarily screening lung tumors for benign and malignant degrees or mixing histopathological images of lung cancer and other cancers to differentiate and classify them, rather than accurately classifying specific subtypes of lung cancer. • Existing methods fail to take full advantage of the vast amount of other data available in modern clinics, and using routinely obtained images of surgical specimen sections for histological classification may be necessary for diagnostic and therapeutic decisions, as innovative tools for clinical data evaluation are needed to augment biopsies and help better characterize the disease, given the complexity of lung cancer classification and the limitations of current practice.…”

Section: Introductionmentioning

confidence: 99%

“…1. Lung cancer classification methods based on low-dose computed tomography (CT) images (Anderson and Davis, 2018), which are mainly studied based on public datasets such as LIDC-IDRI and LUNA16 (Gugulothu and Balaji, 2023;Liu et al, 2023;Zhu et al, 2023). 2.…”

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

KD_ConvNeXt: knowledge distillation-based image classification of lung tumor surgical specimen sections

Zheng,

Yao,

Lin

et al. 2023

Front. Genet.

View full text Add to dashboard Cite

Introduction: Lung cancer is currently among the most prevalent and lethal cancers in the world in terms of incidence and fatality rates. In clinical practice, identifying the specific subtypes of lung cancer is essential in diagnosing and treating lung lesions.Methods: This paper aims to collect histopathological section images of lung tumor surgical specimens to construct a clinical dataset for researching and addressing the classification problem of specific subtypes of lung tumors. Our method proposes a teacher-student network architecture based on a knowledge distillation mechanism for the specific subtype classification of lung tumor histopathological section images to assist clinical applications, namely KD_ConvNeXt. The proposed approach enables the student network (ConvNeXt) to extract knowledge from the intermediate feature layers of the teacher network (Swin Transformer), improving the feature extraction and fitting capabilities of ConvNeXt. Meanwhile, Swin Transformer provides soft labels containing information about the distribution of images in various categories, making the model focused more on the information carried by types with smaller sample sizes while training.Results: This work has designed many experiments on a clinical lung tumor image dataset, and the KD_ConvNeXt achieved a superior classification accuracy of 85.64% and an F1-score of 0.7717 compared with other advanced image classification methods

show abstract