MSFSegNet: A multi-scale feature fusion model for instance segmentation in adult liver ultrasound images
These videos have been matched automatically. Contact us if they are not relevant.
These videos have been matched automatically. Contact us if they are not relevant.
View abstract on PubMed
Summary
This summary is machine-generated.A new AI model, MSFSegNet, accurately segments liver and accessory structures in ultrasound images, improving early disease detection. This automated approach overcomes human error and time constraints in medical imaging analysis.
Area Of Science
- Medical Imaging
- Artificial Intelligence
- Computer-Aided Diagnosis
Background
- Liver diseases often progress undetected due to a lack of early symptoms.
- Manual segmentation of liver and accessory structures (LAS) in ultrasound images is time-consuming and error-prone.
- Accurate segmentation is crucial for timely diagnosis and treatment planning.
Purpose Of The Study
- To introduce MSFSegNet, a novel instance segmentation model for adult liver ultrasound images.
- To address the limitations of manual segmentation in terms of time and accuracy.
- To enhance the early detection of liver diseases through improved image analysis.
Main Methods
- Developed MSFSegNet, an instance segmentation model incorporating a multi-scale feature fusion network (CCMC).
- Integrated an adaptive downsampling method (ODConv) and the Convolutional Block Attention Module (CBAM) for enhanced accuracy.
- Focused on improving segmentation of small anatomical structures within liver ultrasound images.
Main Results
- MSFSegNet achieved high performance metrics: Precision (94.4%), Recall (91.8%), and mAP@0.5 (95.7%) in position evaluation.
- Segmentation tasks yielded excellent results: Precision (93.9%), Recall (91.3%), and mAP@0.5 (94.8%).
- The model significantly outperformed existing methods in accuracy and efficiency.
Conclusions
- MSFSegNet shows significant potential for computer-aided diagnosis in liver ultrasound imaging.
- The model provides a robust solution for accurate segmentation of liver and accessory structures (LAS).
- Future research will focus on computational efficiency and application to pathological cases.
These videos have been matched automatically. Contact us if they are not relevant.
These videos have been matched automatically. Contact us if they are not relevant.
