No abstract

No abstract

The liver is an irreplaceable organ in the human body, maintaining life activities and metabolism. Malignant tumors of the liver have a high mortality rate at present. Computer-aided segmentation of the liver and tumors has significant effects on clinical diagnosis and treatment. There are still many challenges in the segmentation of the liver and liver tumors simultaneously, such as, on the one hand, that convolutional kernels with fixed geometric structures do not match complex, irregularly shaped targets; on the other, pooling during convolution results in a loss of spatial contextual information of images. In this work, we designed a cascaded U-ADenseNet with coarse-to-fine processing for addressing the above issues of fully automatic segmentation. This work contributes multi-resolution input images and multi-layered channel attention combined with atrous spatial pyramid pooling densely connected in the fine segmentation. The proposed model was evaluated by a public dataset of the Liver Tumor Segmentation Challenge (LiTS). Our approach attained competitive liver and tumor segmentation scores that exceeded other methods across a wide range of metrics.

In this paper, we use the embedded theory and wavelet algorithm to improve image coding, and get a new optimization algorithm of track and field image, and establish the wavelet reconstruction mathematical model of image optimization. In order to verify the effectiveness and reliability of the wavelet reconstruction algorithm, this paper uses computer embedded PLC control system and 317-2 PN/DP special CPU to establish the optimization system of track and field path, and debug a program by using Step7 software, which realizes the calculation function of embedded system. Finally, this paper does wavelet reconstruction on real time image, we obtain the reconstruction wavelet optimal path, and output it in the form of digital image, and use the drawing function to realize re-draw function of image, finally we get the route optimization figure of track and field teaching. It provides a new method for computer teaching process.

Automatic and accurate 3D pancreas segmentation based on deep learning, playing an important role in medical professional's diagnosis and treatment of diseases, has received a lot of attention from medical image processing community. Although sufficient works on 3D pancreas segmentation have been conducted in recent years, they embrace two shortcomings: one is that the server environment required for segmentation is too excellent and not friendly to the public, and the other is that the segmentation accuracy needs to be further improved. In response to the above problems, we propose a Selecting the Overlap Method (SOM) operation that effectively relieves the memory pressure and associates contextual information. To accurately segment 3D medical images, a 3D segmentation model called RC-3DUnet is designed to reduce the parameters during segmentation compared with the general network. This integrates our newly proposed coarse-to-fine segmentation method, which utilizes SOM and RC-3DUnet, to both reduce the running memory environment required for 3D segmentation of the pancreas and increase the accuracy rate. Extensive experiments were conducted on the Medical Segmentation Decathlon (MSD) pancreas segmentation and the National Institutes of Health (NIH) pancreas segmentation datasets, showing our results not only showed higher accuracy than other excellent methods but also demonstrated that our method is more lightweight.