Abstract: The extensive use of medical imaging has become a routine practice in modern medical health care centres. It is used almost in every stage of patient management system. However, it is intuition or expertise of physician to choose a modality or alternative modality wisely for managing the patient as single modality has limitations. Therefore, single modality is necessarily ruled out in diagnosis and treatment processes. Multimodality medical image analysis plays a significant role in the diagnosis, treatment planning, delivery of treatment, and review of patient’s response to the treatment. In this thesis an attempt is made to aid radiologist with new fused image created from two modality images for the better visualization and interpretation of abnormalities in context with the purpose of accurate diagnosis, to prepare precise treatment plan, to classify the stages of diseases, and to review the effectiveness of the treatment. The proposed research work presents the feature based fusion algorithms in wavelet domain to combine the relevant and complementary spectral features of two modalities namely computed tomography (CT) and magnetic resonance imaging (MRI). The directional features of source modality images are extracted using various proposed wavelet transforms viz. nonsubsampled rotated wavelet transform, nonsubsampled rotated dual tree complex wavelet transform, dual tree complex wavelet packet transform, M-Band wavelet & MBand complex wavelet transforms, and rotated Daubechies complex wavelet transform. These spectral features are combined in new composite space using appropriate fusion rules. The corresponding inverse transform is used to create fused image.
The multimodality medical image fusion is a powerful technique for analysis of lesions. The fusion approaches presented in this thesis are used for three main applications i.e. radiotherapy (RT), diagnosis and stage identification of neurocysticercosis (NCC), and the disease management in hepatocellular carcinoma (HCC). The fused images are useful in radiotherapy for accurate localization of tumour, to visualize its complete spread, and precise treatment plan for the target volumes without affecting critical organs. In second application, fused images help radiologists to identify the stage disease of neurocysticercosis. It is beneficial to finalize the treatment plan as per the stage disease. In HCC, the fusion results are presenting the level of damage to liver, accurate localization of masses, classify the disease, and confirming diagnosis which is helpful in conclusive treatment plan. Fusion process is also useful in the post treatment follow ups of neurocysticercosis.
Keywords: Medical, imaging, image processing technique
|
DOI:
10.17148/IJARCCE.2022.114185
