Systems and methods for medical image analysis and classification
Systems and methods for performing image classification are disclosed. The methods include receiving a plurality of magnetic resonance imaging (MM) images that each include metadata. The plurality of images are sorted into one or more groups using the metadata. The methods further include, for each of the one or more groups: identifying a subset of images, generating for a classification label for each image in the subset of images using a classifier, identifying a first classification label that is associated with a maximum number of images in the subset of images, and assigning the first classification label to each image in that group.
1 . A method of performing image classification, the method comprising, by a processor:
receiving a plurality of magnetic resonance imaging (MRI) images, each of the plurality of images comprising metadata;
sorting, using the metadata, the plurality of images into one or more groups; and
for each of the one or more groups:
identifying a subset of images;
generating, using a classifier, for each image in the subset of images, a classification label;
identifying a first classification label, the first classification label being associated with a maximum number of images in the subset of images; and
assigning the first classification label to each image in that group;
wherein identifying the subset of images comprises selecting images that lie, based on a corresponding patient orientation, within a threshold pixel distance of an imaging plane corresponding to the group of images.
2 . The method of claim 1 , wherein the classification label comprises at least one of the following: a sequence type label, an anatomy label, or a view type label.
3 . The method of claim 1 , wherein the metadata is retrieved from a DICOM header of that image.
4 . The method of claim 1 , wherein sorting the plurality of images into one or more groups comprises sorting the images such that each image in a group includes an attribute value that is either similar to corresponding attribute values of other images in the group or within a threshold of the corresponding attribute values.
5 . The method of claim 4 , wherein the metadata comprises one or more of the following attributes: patient attributes, pre-clinical or clinical study attributes, MRI scanner identification, series instance UID, series number, series description, protocol name, series time of acquisition, image pixels, image plane, weighting classification, time of acquisition of image, imaged organ identification, disease model, patient orientation, view type, slice thickness, or bolus magnetic contrast agent bolus identity.
6 . The method of claim 4 , wherein identifying the subset of images comprises:
determining a median attribute value of an attribute for images included in that group; and
selecting images that are within a threshold distance of the median attribute value.
7 . The method of claim 4 , further comprising selecting an attribute for sorting the plurality of images into one or more groups, the selected attribute being configured to maximize a number of the one or more groups.
8 . The method of claim 1 , wherein the subset of images comprises an odd number of images.
9 . The method of claim 1 , further comprising identifying and discarding, one or more of the plurality of images that are in a format that is not DICOM compatible.
10 . The method of claim 1 , further comprising storing, in a data representation, the first classification label in association with each image in that group.
11 . A system for performing image classification, the system comprising:
a processor; and
a non-transitory computer readable medium comprising programming instructions that when executed by the processor will cause the processor to:
receive a plurality of magnetic resonance imaging (MRI) images, each of the plurality of images comprising metadata;
sort, using the metadata, the plurality of images into one or more groups; and
for each of the one or more groups:
identify a subset of images,
generate, using a classifier, for each image in the subset of images, a classification label,
identify a first classification label, the first classification label being associated with a maximum number of images in the subset of images; and
assign the first classification label to each image in that group;
wherein identifying the subset of images comprises selecting images that lie, based on a corresponding patient orientation, within a threshold pixel distance of an imaging plane corresponding to the group of images.
12 . The system of claim 11 , wherein the classification label comprises at least one of the following: a sequence type label, an anatomy label, or a view type label.
13 . The system of claim 11 , wherein the metadata is retrieved from a DICOM header of that image.
14 . The system of claim 11 , wherein the instructions that cause the processor to sort the plurality of images into one or more groups comprise instructions to cause the processor to sort the images such that each image in a group includes an attribute value that is either similar to corresponding attribute values of other images in the group or within a threshold of the corresponding attribute values.
15 . The system of claim 14 , wherein the metadata comprises one or more of the following attributes: patient attributes, pre-clinical or clinical study attributes, MRI scanner identification, series instance UID, series number, series description, protocol name, series time of acquisition, image pixels, image plane, weighting classification, time of acquisition of image, imaged organ identification, disease model, patient orientation, view type, slice thickness, or bolus magnetic contrast agent bolus identity.
16 . The system of claim 14 , wherein the instructions that cause the processor to identify the subset of images comprise instructions to cause the processor to:
determine a median attribute value of an attribute for images included in that group; and
select images that are within a threshold distance of the median attribute value.
17 . The system of claim 14 , further comprising instructions to cause the processor to select an attribute for sorting the plurality of images into one or more groups, the selected attribute being configured to maximize a number of the one or more groups.
18 . The system of claim 11 , further comprising instructions to cause the processor to identify and discard, one or more of the plurality of images that are in a format that is not DICOM compatible.
19 . The system of claim 11 , further comprising instructions to cause the processor to store, in a data representation, the first classification label in association with each image in that group.