Internal evaluation of registration results for radiographic images

This wort focuses on internal gray level based evaluation of imageregistration results. The motivation is to provide all approach forself-diagnosis in the scope of a patient alignment system based onrigid registration of real and reconstructed X-ray images. As anautomatic system should provide expressive indicators for the correctnessof the outcome, we propose a method to estimate the probability forthe resulting transformations to lie within a predefined window ofacceptable values. Based purely on image gray values, the approachis independent from previous knowledge about the images. By registrationof corresponding fragments of both images we generate redundancyand define the probability density of the resulting transformations.The proposed method is tested comparing digital reconstructed radiographs(DRRs) to X-ray images. By introducing geometric and radiometricdeviations we show that a reliable self-diagnosis is possible.     «

This wort focuses on internal gray level based evaluation of imageregistration results. The motivation is to provide all approach forself-diagnosis in the scope of a patient alignment system based onrigid registration of real and reconstructed X-ray images. As anautomatic system should provide expressive indicators for the correctnessof the outcome, we propose a method to estimate the probability forthe resulting transformations to lie within a predefined window ofacceptable values. Based purely...     »