Semi-supervised learning leveraging cross-domain data for medical imaging analysis
Systems and methods for performing a medical imaging analysis task are provided. An input medical image in a first modality is received. Features are extracted from the input medical image using a first machine learning based encoding network. A medical imaging analysis task is performed on the input medical image based on the extracted features by, in one embodiment, decoding the extracted features to generate results of the medical imaging analysis task using a machine learning based decoding network. Results of the medical imaging analysis task are output. In one embodiment, the first machine learning based encoding network is jointly trained with a second machine learning based encoding network with an unsupervised loss using unannotated pairs of training images. Each of the unannotated pairs comprise a first training image in the first modality and a second training image in a second modality. In one embodiment, the first machine learning based encoding network is also jointly trained with the machine learning based decoding network with a supervised loss using annotated training images.
1. A computer-implemented method comprising:
receiving an input medical image in a first modality;
extracting features from the input medical image using a first machine learning based encoding network, wherein the first machine learning based encoding network is jointly trained with a second machine learning based encoding network using unannotated pairs of training images, each of the unannotated pairs comprising a first training image in the first modality and a second training image in a second modality;
performing a medical imaging analysis task on the input medical image based on the extracted features; and
outputting results of the medical imaging analysis task.
2. The computer-implemented method of claim 1 , wherein the first machine learning based encoding network is jointly trained with the second machine learning based encoding network using similar pairs and dissimilar pairs of the unannotated pairs of training images.
3. The computer-implemented method of claim 2 , wherein the similar pairs are determined to be similar based on location and time.
4. The computer-implemented method of claim 3 , wherein the location and the time are extracted from a header of the similar pairs of training images.
5. The computer-implemented method of claim 2 , wherein the first machine learning based encoding network is jointly trained with the second machine learning based encoding network with an unsupervised loss to minimize a distance between features of the similar pairs and to maximize a distance between features of the dissimilar pairs.
6. The computer-implemented method of claim 1 , wherein performing a medical imaging analysis task on the input medical image based on the extracted features comprises:
decoding the extracted features to generate the results of the medical imaging analysis task using a machine learning based decoding network.
7. The computer-implemented method of claim 6 , wherein the first machine learning based encoding network is jointly trained with the machine learning based decoding network with a supervised loss using annotated training images.
8. The computer-implemented method of claim 1 , wherein the input medical image depicts an anatomical object of a patient and performing a medical imaging analysis task on the input medical image based on the extracted features comprises:
segmenting the anatomical object from the input medical image.
9. The computer-implemented method of claim 1 , wherein the first modality comprises one of cine magnetic resonance imaging (MRI), T1 mapping, or T2 mapping and the second modality comprises a different one of the cine MRI, the T1 mapping, or the T2 mapping.
10. An apparatus comprising:
means for receiving an input medical image in a first modality;
means for extracting features from the input medical image using a first machine learning based encoding network, wherein the first machine learning based encoding network is jointly trained with a second machine learning based encoding network using unannotated pairs of training images, each of the unannotated pairs comprising a first training image in the first modality and a second training image in a second modality;
means for performing a medical imaging analysis task on the input medical image based on the extracted features; and
means for outputting results of the medical imaging analysis task.
11. The apparatus of claim 10 , wherein the first machine learning based encoding network is jointly trained with the second machine learning based encoding network using similar pairs and dissimilar pairs of the unannotated pairs of training images.
12. The apparatus of claim 11 , wherein the similar pairs are determined to be similar based on location and time.
13. The apparatus of claim 12 , wherein the location and the time are extracted from a header of the similar pairs of training images.
14. The apparatus of claim 11 , wherein the first machine learning based encoding network is jointly trained with the second machine learning based encoding network with an unsupervised loss to minimize a distance between features of the similar pairs and to maximize features of the dissimilar pairs.
15. A non-transitory computer readable medium storing computer program instructions, the computer program instructions when executed by a processor cause the processor to perform operations comprising:
receiving an input medical image in a first modality;
extracting features from the input medical image using a first machine learning based encoding network, wherein the first machine learning based encoding network is jointly trained with a second machine learning based encoding network using unannotated pairs of training images, each of the unannotated pairs comprising a first training image in the first modality and a second training image in a second modality;
performing a medical imaging analysis task on the input medical image based on the extracted features; and
outputting results of the medical imaging analysis task.
16. The non-transitory computer readable medium of claim 15 , wherein the first machine learning based encoding network is jointly trained with the second machine learning based encoding network using similar pairs and dissimilar pairs of the unannotated pairs of training images.
17. The non-transitory computer readable medium of claim 15 , wherein performing a medical imaging analysis task on the input medical image based on the extracted features comprises:
decoding the extracted features to generate the results of the medical imaging analysis task using a machine learning based decoding network.
18. The non-transitory computer readable medium of claim 17 , wherein the first machine learning based encoding network is jointly trained with the machine learning based decoding network with a supervised loss using annotated training images.
19. The non-transitory computer readable medium of claim 15 , wherein the input medical image depicts an anatomical object of a patient and performing a medical imaging analysis task on the input medical image based on the extracted features comprises:
segmenting the anatomical object from the input medical image.
20. The non-transitory computer readable medium of claim 15 , wherein the first modality comprises one of cine magnetic resonance imaging (MRI), T1 mapping, or T2 mapping and the second modality comprises a different one of the cine MRI, the T1 mapping, or the T2 mapping.