System and method for automated labeling and annotating unstructured medical datasets
Supervised and unsupervised learning schemes may be used to automatically label medical images for use in deep learning applications. Large labeled datasets may be generated from a small initial training set using an iterative snowball sampling scheme. A machine learning powered automatic organ classifier for imaging datasets, such as CT datasets, with a deep convolutional neural network (CNN) followed by an organ dose calculation is also provided. This technique can be used for patient-specific organ dose estimation since the locations and sizes of organs for each patient can be calculated independently.
1. A method for automatically processing unstructured medical imaging data to generate organ classified images, comprising:
a) acquiring medical image data of a subject;
b) subjecting the medical image data of the subject to a neural network to generate classified image data;
c) segmenting the classified image data using the convolutional neural network to generate segmented images that distinguish between at least two different organs in the classified regions in the images;
d) generating a report of a calculated radiation dose for at least one of the at least two different organs in the segmented images; and
further comprising measuring an effective diameter for a cross section of the subject based upon the images.
2. The method of claim 1 wherein the calculated radiation dose for the at least one of the at least two different organs is based upon organ volume and shape determined by the effective diameter.
3. The method of claim 1 wherein the report includes the ratio of normalized organ dose by CTDI vol to the patient effective diameter.
4. The method of claim 1 wherein classified image data includes at least one of identifying a brain, eye, nose, salivary gland, thyroid, upper lung, thymus, heart, chest, abdomen, pelvis, urinary bladder, genitals, or a Leg.
5. A method for automatically processing unstructured medical imaging data to generate organ classified images, comprising:
a) acquiring medical image data of a subject;
b) subjecting the medical image data of the subject to a neural network to generate classified image data;
c) segmenting the classified image data using the convolutional neural network to geberate segmented images that distinguish between at least two different organs in the classified regions in the images;
d) generating a report of a calculated radiation dose for at least one of the at least two different organs in the segmented images; and
further comprising calculating a correction factor for generating the report of the calculated radiation dose, wherein the correction factor is calculated at each organ using data provided by two different manufacturers.
6. A method for organ classification for unstructured medical datasets, comprising:
a) acquiring images of a region of a subject and labeling the images to generate a training dataset with the images;
b) training a convolutional neural network with the training dataset;
c) classifying a region in the images using the trained network;
d) segmenting the classified images using the convolutional neural network to generate segmented images that distinguish between at least two different organs in the classified regions in the images;
e) generating a report of a calculated radiation dose for at least one of the at least two different organs in the segmented images; and
further comprising measuring an effective diameter for a cross section of the subject based upon the images.
7. The method of claim 6 wherein the calculated radiation dose for the at least one of the at least two different organs is based upon organ volume and shape determined by the effective diameter.
8. The method of claim 6 wherein the report includes the ratio of normalized organ dose by CTDI vol to the patient effective diameter.
9. The method of claim 6 wherein classifying a region includes at least one of identifying a brain, eye, nose, salivary gland, thyroid, upper lung, thymus, heart, chest, abdomen, pelvis, urinary bladder, genitals, or a Leg.
10. A method for organ classification for unstructured medical datasets, comprising:
a) acquiring images of a region of a subject and labeling the images to generate a training dataset with the images;
b) training a convolutional neural network with the training dataset;
c) classifying a region in the images using the trained network;
d) segmenting the classified images using the convolutional neural network to generate segmented images that distinguish betweeen at least two different organs in the classified regions in the images;
e) generating a report of a claculated radiation does for at least one of the at least two different organs in the segmented images; and
further comprising calculating a correction factor for generating the report of the calculated radiation dose, wherein the correction factor is calculated at each organ using data provided by two different manufacturers.
11. A system for organ classification for unstructured medical datasets, comprising:
a computer system configured to:
i) acquire images of a region of a subject and labeling the images to generate a training dataset with the images;
ii) train a convolutional neural network with the training dataset;
iii) classify a region in the images using the trained network;
iv) segment the classified images using the convolutional neural network to generate segmented images that distinguish between at least two different organs in the classified regions in the images;
v) generate a report of a calculated radiation dose for at least one of the at least two different organs in the segmented images; and
further comprising calculating a correction factor for generating the report of the calculated radiation dose, wherein the correction factor is calculated at each organ using data provided by two different manufacturers.
12. The system of claim 11 wherein classify a region includes at least one of identifying a brain, eye, nose, salivary gland, thyroid, upper lung, thymus, heart, chest, abdomen, pelvis, urinary bladder, genitals, or a Leg.
13. A system for organ classification for unstructured medical datasets, comprising:
a computer system configured to:
i) acquire images of a region of a subject and labeling the images to generate a training dataset with the images;
ii) train a convolutional neural network with the training dataset;
iii) classify a region in the images using the trained network;
iv) segment the classified images using the convolutional neural network to generate segmented images that distinguish between at least two different organs in the classified regions in the images;
v) generate a report of a calculated radiation dose for at least one of the at least two different organs in the segmented images; and
further comprising measuring an effective diameter for a cross section of the subject based upon the images.
14. The system of claim 13 wherein the calculated radiation dose for the at least one of the at least two different organs is based upon organ volume and shape determined by the effective diameter.
15. The system of claim 13 wherein the report includes the ratio of normalized organ dose by CTDI vol to the patient effective diameter.
16. A method for automatically processing unstructured medical imaging data to generate organ classified images, comprising:
a) acquiring medical image data of a subject;
b) subjecting the medical image data of the subject to a neural network to generate classified image data;
c) segmenting the classified image data using the convolutional neural network to generate segmented images that distinguish between at least two different organs in the classified regions in the images;
d) generating a report of a claculated radiation does for at least one of the at least two different organs in the segmented images; and
calculating the calculated radiation dose using a correction factor, wherein the correction factor represents a correlation between a ratio of a normalized organ dose for the at least one of the at least two different organs and a patient effective diameter.