Articulatory Motion Synthesis System Based On Ultrasound And Magnetic Resonance Images

With the continuous development of information technology,communication between human beings has become increasingly close.Language communication is the main way of human communication and plays an important role in human economic,cultural and work.However,to date,“language barriers”remain the main problem encountered in daily human communication.“Language barrier” usually includes the barriers due to geographical and cultural differences and the barriers due to damage to the vocal organs.Traditional speech synthesis systems focus more on the processing of speech signals.However,researchers need a framework that can “visualize” data analysis of the articulatory movement process.Based on this,researchers focus on the study of articulatory synthesis systems.By definition,the articulatory motion synthesis system is a technique for reproducing the motion by mechanical and electronic means.To date,the techniques for this system include ultrasound and magnetic resonance images processing techniques.Ultrasound has good directionality and penetration,and it is often used for scanning human organs.But it has the disadvantages of low contrast,poor definition and strong pixel interference.In contrast,magnetic resonance images have high definition and complete imaging.However,it is difficult to obtain real-time images.Both of these technologies have their own pronunciation synthesis system.Due to their shortcomings,it is difficult to fully play a role in the visual study of articulatory movement.In order to solve this problem,the work proposes a pronunciation motion synthesis system based on ultrasound and magnetic resonance images.The work combines the advantages of ultrasound images with magnetic resonance images.The system uses an ultrasonic collector to collect position change information of the joint nodes of the pronunciation part in real time.At the same time,an automated method is used to build a magnetic resonance image library.The ultrasound image is then mapped in real time to the data set in the image library.Finally,the results are fed back in a dynamic manner.The innovation of the work is to propose a method to map two different types of tomographic images in real time.And the premises of the correct rate of mapping are the accurately extracting outline method and theautomated establishment of the magnetic resonance image library.The design and implementation of this subject provides a visual method for studying the synthesis of articulatory motion.