An accepted practice among radiologists to detect anomalous structures or architectural deformation in mammography images is comparing images of the same patient.
There exist three ways of comparison:
Temporal: comparing mammography images for the same breast at different screening examinations.
Contralateral: comparing the mammography images of left and right breasts.
Ipsilateral: comparing the craniocaudal (CC) and mediolateral oblique (MLO) views of the same breast.
An accepted practice among radiologists to detect anomalous structures or architectural deformation in mammography images is comparing images of the same patient.
There exist three ways of comparison:
Temporal: comparing mammography images for the same breast at different screening examinations.
Contralateral: comparing the mammography images of left and right breasts.
Ipsilateral: comparing the craniocaudal (CC) and mediolateral oblique (MLO) views of the same breast.
Comparing mammography images has never been a trivial task since appearance of an image can be altered because of several technical and natural factors such as
• Choice of image acquisition parameters
• Positioning and compression of the breast
• Changes in breast anatomy
Mammography image registration (alignment) is a technique for suppressing technical variations and retain genuine changes in the breast. It enables radiologists to efficiently compare two images and avoid nontarget discrepancies.
The comparison module of WisdomX platform employs a propriety mammography images registration technique that is enable to model transformations by means of deep neural networks.
The comparison module of WisdomX platform employs a propriety mammography images registration technique that is enable to model transformations by means of deep neural networks.
