Method and system for detecting chest x-ray thoracic diseases utilizing multi-view multi-scale learning

A method for biomedical image analysis, comprising:

• receiving, at a communication interface, a plurality of unannotated biomedical images, the biomedical images including a first image and a second image being acquired in a pair by an image acquisition device;
• determining, by at least one processor, that the first image is in a first view and the second image is in a second view;
• assigning, by the at least one processor, the first image of the plurality of unannotated biomedical images to a first processing path comprising a first set of processing modules designed for the first orientation;
• assigning, by the at least one processor, the second image of the plurality of unannotated biomedical images to a second processing path comprising a second set of processing modules designed for the second view;
• processing, by the at least one processor, the first image in the first processing path in parallel with processing the second image in the second processing path, wherein the first set of processing modules share processing parameters with the second set of processing modules;
• and
• providing, by the at least one processor, a diagnostic output based on the processing of the first image and the second image.

A non-transitory computer-readable medium having a computer program stored thereon, wherein the computer program, when executed by at least one processor, performs a method for biomedical image analysis, the method comprising:

• receiving a plurality of unannotated biomedical images, the biomedical images including a first image and a second image being acquired in a pair by an image acquisition device;
• and
• determining that the first image is in a first view and the second image is in a second view;
• assigning the first image of the plurality of unannotated biomedical images to a first processing path comprising a first set of processing modules designed for the first orientation;
• assigning the second image of the plurality of unannotated biomedical images to a second processing path comprising a second set of processing modules designed for the second view;
• processing the first image in the first processing path in parallel with processing the second image in the second processing path, wherein the first set of processing modules share processing parameters with the second set of processing modules;
• and
• providing a diagnostic output based on the processing of the first image and the second image.