Abstract
Background: Accurate segmentation of liver tumor regions in medical images is of great significance for clinical diagnosis and the planning of surgical treatments. Recent advancements in machine learning have shown that convolutional neural networks are powerful in such image processing while largely reducing human labor. However, the variable shape, fuzzy boundary, and discontinuous tumor region of liver tumors in medical images bring great challenges to accurate segmentation. The feature extraction capability of a neural network can be improved by expanding its architecture, but it inevitably demands more computing resources in training and hyperparameter tuning.
Methods: This study presents a Dynamic Context Encoder Network (DCE-Net), which incorporates multiple new modules, such as the Involution Layer, Dynamic Residual Module, Context Extraction Module, and Channel Attention Gates, for feature extraction and enhancement.
Results: In the experiment, we used a liver tumor CT dataset of LiTS2017 to train and test the DCE-Net for liver tumor segmentation. The experimental results showed that the four evaluation indexes of the method, precision, recall, dice, and AUC, were 0.8961, 0.9711, 0.9270, and 0.9875, respectively. Furthermore, our ablation study reported that the accuracy and training efficiency of our network were markedly superior to the networks without involution or dynamic residual modules.
Conclusion: Therefore, the DCE-Net proposed in this study has great potential for automatic segmentation of liver lesion tumors in the clinical diagnostic environment.
Keywords: Dynamic context encoder network (DCE-Net), Liver tumor, Segmentation, Feature extraction, Attention, Liver cancer.
Call for Papers in Thematic Issues
Advancements and Applications of Photon Counting CT: Transforming Imaging Precision and Quantification
Photon Counting Computed Tomography (CT) represents a transformative advancement in medical imaging technology, offering unprecedented levels of precision and diagnostic capability. This special issue will explore the latest innovations and applications of photon counting CT, highlighting its advantages over traditional energy-integrating detectors. The scope includes an in-depth examination of its ...read more
Advancements in Deep Learning based Medical Image and Signal Processing for Healthcare Applications
The accurate predication from medical data (signals and images) analytics will be achieved through the use of cutting-edge AI tools for tailoring medical treatment to each patient. When the agents are done analysing the user's health data, the user can consult with the agents to receive advice and the most ...read more
AI-Driven Diagnostics in Radiology and Medical Imaging
Medical imaging data, such as X-rays, MRIs, ultrasounds, CT scans, and DXAs, can be analysed by AI algorithms to help healthcare professionals identify and diagnose diseases more rapidly and accurately. Among the most cutting-edge applications of artificial intelligence (AI) in healthcare are imaging-related applications like cancer detection. AI could revolutionise ...read more
Applying medical image processing techniques for early cancer detection in mammograms
Breast cancer survival and successful treatment are significantly boosted by early identification. Breast cancer screening is most commonly accomplished with mammography. The radiologist's experience and the image quality, among several other factors, might affect how well a mammogram is interpreted. Enhancing mammography analysis and increasing cancer early detection are made ...read more
14