Analysis of segmentation of lung parenchyma based on deep learning methods

Abstract

Precise segmentation of lung parenchyma is essential for effective analysis of the lung. Due to the obvious contrast and large regional area compared to other tissues in the chest, lung tissue is less difficult to segment. Special attention to details of lung segmentation is also needed. To improve the quality and speed of segmentation of lung parenchyma based on computed tomography (CT) or computed tomography angiography (CTA) images, the 4th International Symposium on Image Computing and Digital Medicine (ISICDM 2020) provides interesting and valuable research ideas and approaches. For the work of lung parenchyma segmentation, 9 of the 12 participating teams used the U-Net network or its modified forms, and others used the methods to improve the segmentation accuracy include attention mechanism, multi-scale feature information fusion. Among them, U-Net achieves the best results including that the final dice coefficient of CT segmentation is 0.991 and the final dice coefficient of CTA segmentation is 0.984. In addition, attention U-Net and nnU-Net network also performs well. In this review paper, the methods chosen by 12 teams from different research groups are evaluated and their segmentation results are analyzed for the study and references to those involved.

A Contrastive Predictive Coding-Based Classification Framework for Healthcare Sensor Data

“…Figure 6 shows the change trend of the accuracy of the classification model prediction when the pretext task is used and the supervised learning is used when samples of different proportions are used. e supervised learning here refers to the model without using the encoder part for i < 32 do (4) for j > 16 do (5) Random select 16 numbers from the total categories. Each number is repeated 8 times and the shape is (16,8).…”

“…(6) Fill train label batch as 1 (7) end for (8) for k in (16, 32) do (9) Random select 16 numbers from the total categories. Each number is repeated 4 times and the shape is (16,4). is is the lower left part of Figure 1.…”

“…(10) Select 16 numbers different from the previously selected category. Each number is repeated 4 times and the shape is (16,4). is is the lower right part of Figure 1.…”

“…This no longer meets the requirements of sustainable urban development, prompting a shift from hospital-centered to family-centered healthcare [ 2 ]. The application of various deep learning algorithms has made it less difficult to automatically classify diseases and has greatly improved the accuracy of disease classification [ 3 , 4 ]. The classification model can be paired with various IoT devices for real-time diagnosis [ 5 ], and patients can grasp their health status at home without having to go to the hospital for checkups every time, which will ease the tension on medical resources and help the construction of smart medical care to achieve sustainable urban development.…”

Deep Learning-Based Text Emotion Analysis for Legal Anomie

“…Convolutional neural network (CNN; Tan et al, 2021) is an important feature extraction model for psychological text. Because of its strong local feature extraction ability, it has achieved good results in the field of text classification.…”