Integration of Genetic Programming and TABU Search Mechanism for Automatic Detection of Magnetic Resonance Imaging in Cervical Spondylosis

Cervical spondylosis is a kind of degenerative disease which not only occurs in elder patients. The age distribution of patients is unfortunately decreasing gradually. Magnetic Resonance Imaging (MRI) is the best tool to confirm the cervical spondylosis severity but it requires radiologist to spend a lot of time for image check and interpretation. In this study, we proposed a prediction model to evaluate the cervical spine condition of patients by using MRI data. Furthermore, to ensure the computing efficiency of the proposed model, we adopted a heuristic programming, genetic programming (GP), to build the core of refereeing engine by combining the TABU search (TS) with the evolutionary GP. Finally, to validate the accuracy of the proposed model, we implemented experiments and compared our prediction results with radiologist’s diagnosis to the same MRI image. The experiment found that using clinical indicators to optimize the TABU list in GP+TABU got better fitness than the other two methods and the accuracy rate of our proposed model can achieve 88% on average. We expected the proposed model can help radiologists reduce the interpretation effort and improve the relationship between doctors and patients.