研究目的
To develop a novel computer-aided diagnosis (CAD) framework for accurate classification of Fatty Liver Disease (FLD) using ultrasound images of the liver.
研究成果
The proposed CAD framework using deep learning and transfer learning achieved a classification accuracy of 90.6% for FLD in ultrasound images, demonstrating potential to aid clinicians in accurate diagnosis.
研究不足
The study is limited by the size of the dataset and the computational resources required for training deep learning models. Future work includes porting the algorithm to a hardware platform for more translational applications.
1：Experimental Design and Method Selection:
The study proposes a CAD framework using convolution neural networks and transfer learning with a pre-trained VGG-16 model.
2：Sample Selection and Data Sources:
The database consists of 81 normal liver images and 76 fatty liver images collected from a Siemens Acuson S1000 ultrasound scanner.
3：List of Experimental Equipment and Materials:
Siemens Acuson S1000 ultrasound scanner with curved array transducer.
4：Experimental Procedures and Operational Workflow:
Images were cropped to remove non-diagnostic information, resized to 224×224 pixels, and augmented to increase the dataset size. The VGG-16 model was fine-tuned for the classification task.
5：Data Analysis Methods:
Performance was analyzed using classification accuracy, confusion matrix, Fscore, Precision, and Recall.
独家科研数据包，助您复现前沿成果，加速创新突破
获取完整内容
