Identification of defects in imaging scans

1. A method of reviewing neural scans, comprising:

receiving at least one landmark corresponding to an anatomical region;

receiving a plurality of images of tissue including the anatomical region;

generating a neural network configured to differentiate between healthy tissue and unhealthy tissue within the anatomical region, wherein the neural network is generated by a machine learning process configured to receive the plurality of images of tissue and generate a plurality of weighting factors configured to differentiate between healthy tissue and unhealthy tissue;

receiving at least one patient image of tissue including the anatomical region; and

determining, via the neural network, whether the at least one patient image of tissue includes healthy or unhealthy tissue.

2. The method of claim 1 , wherein the neural network includes a first set and a second set of identical nodes, wherein the anatomical region comprises a symmetrical anatomical region, and wherein the at least one patient image includes a first patch corresponding to a first side of the symmetrical anatomical region and a second patch corresponding to a second side of the symmetrical anatomical region.

3. The method of claim 2 , wherein the first patch is provided to the first set of identical nodes and the second patch is provided to the second set of identical nodes.

4. The method of claim 2 , wherein the neural network comprises a third set of identical nodes, wherein the third set of identical nodes is configured to receive a reference patch.

5. The method of claim 1 , wherein the network includes a plurality of skip connections configured to minimize overfitting.

6. The method of claim 1 , comprising generating, via the neural network, impact maps configured to identify an influence of one or more voxels in the determination.

7. The method of claim 1 , wherein the plurality of images and the at least one patient image are contrast computed tomography images.

8. The method of claim 1 , comprising:

receiving a plurality of patient images;

determining, via the neural network, a severity of unhealthy tissue in each of the plurality of patient images;

ranking, by the neural network, each of the plurality of patient images according to the determined severity of unhealthy tissue; and

providing each of the plurality of patient images in rank order for further review.

9. The method of claim 1 , wherein the neural network is configured to identify at least one of a hemorrhage, an acute infarct, hydrocephalus, a mass effect, and/or a mass lesion.

10. The method of claim 1 , wherein the neural network is configured to identify unhealthy tissue based on a change in a position of at least one landmark.

11. The method of claim 1 , wherein the neural network is a supervised learning network.

12. A system for reviewing neural scans, comprising:

an imaging modality configured to obtain a neural image of a patient; and

a processor configured to implement a neural network, wherein the neural network is generated by:

receiving at least one landmark corresponding to an anatomical region;

receiving a plurality of images of tissue including the anatomical region; and

performing a machine learning process configured to review the plurality of images of tissue and generate a plurality of weighting factors configured to differentiate between healthy tissue and unhealthy tissue;

wherein the processor is configured to identify, via the neural network, a defect in the neural image.

13. The system of claim 12 , wherein the neural network includes a first set and a second set of identical nodes, wherein the neural scan includes a first patch corresponding to a first side of a symmetrical anatomical region and a second patch corresponding to a second side of the symmetrical anatomical region.

14. The system of claim 13 , wherein the first patch is provided to the first set of identical nodes and the second patch is provided to the second set of identical nodes.

15. The system of claim 13 , wherein the neural network comprises a third set of identical nodes, wherein the third set of identical nodes is configured to receive a reference patch.

16. The system of claim 12 , comprising generating, via the neural network, impact maps configured to identify an influence of one or more voxels.

17. The system of claim 12 , wherein the imaging modality is a computerized-tomography (CT) modality.

18. The system of claim 12 , comprising:

receiving a plurality of patient images;

determining, via the neural network, a severity of unhealthy tissue in each of the plurality of patient images; and

ranking, by the neural network, each of the plurality of patient images according to the determined severity of unhealthy tissue.

19. The system of claim 12 , wherein the defect is at least one of a hemorrhage, an acute infarct, hydrocephalus, a mass effect, and/or a mass lesion.

20. A non-transitory computer-readable medium encoded with computer executable instructions, the computer executable instructions, when executed by a computer in a system for reviewing neural scans, cause the system for reviewing neural scans to execute the steps of:

receiving at least one landmark corresponding to an anatomical region;

receiving a plurality of images of tissue including the anatomical region;

generating a neural network configured to differentiate between healthy tissue and unhealthy tissue within the anatomical region, wherein the neural network is generated by a machine learning process configured to receive the plurality of images of tissue and generate a plurality of weighting factors configured to differentiate between healthy tissue and unhealthy tissue;

receiving at least one patient image of tissue including the anatomical region; and

determining, via the neural network, whether the at least one patient image of tissue includes healthy or unhealthy tissue.