Prediction system, method and computer program product thereof
AssigneeChina Medical University Hospital
InventorsChia-Hung Kao, Shang-Wen Chen, Wei-Chih Shen, Kuo-Chen Wu
Life of this patent
Estimating the term…
Abstract
A prediction system of tumor image-aided prediction of prognosis for patients with uterine cervical cancer is provided. The prediction system includes a data augmentation module and a deep convolution neural network model. The data augmentation module is used to apply a data expansion process of the image data, so as to generate a plurality of slice sets of the uterine cervical cancer tumor. The deep convolution neural network model is used to apply a feature analysis to the slices, so as to predict the prognosis of a patient after chemoradiotherapy.
Claims9
A prediction system used to analyze image data of a uterine cervical tumor, comprising:
- a data augmentation module, performing a data expansion process on the image data so as to generate a plurality of slice sets of the image data;
- and
- an analysis module, performing a feature analysis on the plurality of slice sets through a deep convolution neural network model to predict a prognosis after the treatment;
- wherein the data augmentation module performs an image interpolation on the image data, wherein the image interpolation sets an SUVmax pixel of the image data as a center to form a basis slice set, and the basis slice set is rotated in a specific direction to form a plurality of expanded slice sets.
A prediction method, used to analyze image data of a uterine cervical tumor, the method is performed by a prediction system, and the method comprises steps of:
- performing a data expansion process on the image data through a data augmentation module so as to generate a plurality of slice sets;
- and
- performing a feature analysis on the plurality of slice sets through a deep convolution neural network model to predict a prognosis;
- wherein the data augmentation module performs an image interpolation on the image data, wherein the image interpolation sets an SUVmax pixel of the image data as a center to form a basis slice set, and the basis slice set is rotated in a specific direction to form a plurality of expanded slices.
A computer program product stored in a non-transitory computer-readable medium for the operation of a prediction system, wherein the prediction system is used to analyze image data of a uterine cervical tumor, wherein the computer program product comprises:
- an instruction, enabling the prediction system to perform a data expansion process on the image data so as to generate a plurality of slice sets;
- and
- an instruction, enabling the prediction system to perform a feature analysis on the slice sets so as to predict a prognosis of the person after the treatment;
- wherein the data expansion process comprises steps of:
- performing an image interpolation on the image data, wherein the image interpolation sets an SUVmax pixel of the image data as a center to form a basis slice set, and the basis slice set is rotated in a specific direction to form a plurality of expanded slices.
Dependencies are inferred from the claim text.