A Hybrid Deep Learning Architecture for Medical Ultrasound Images Enhancement in Liver Tumor Diagnosis

Nowadays, medical ultrasound images is required to classifying different types of liver tumor. Therefore, the doctor must be able to determine accurately diagnosis and evaluation of liver tumor diagnosis. The main problem in a liver tumor disease is in determining the correct classification of tumor. Thus, an important challenge for researcher is to produce systems that can continually improve different types of liver tumor pictured in a given image. Therefore, this paper develops a hybrid deep learning architecture with harmony search algorithm for medical ultrasound images in liver tumor diseases in order to classifying different types of liver tumor pictured in a given image and improved accurate diagnosis and evaluation of liver tumor diagnosis. The proposed method is evaluated by experimenting a number of liver images obtained from NCTRlcdb and also from Radiopaedia portal and the results shows better and more effective analysis and provide better and more effective identification.