Temperature-feedback Photothermal Control System And Intelligent Diagnosis Of Tissue Damage

The photothermal relationship between laser and biological tissue has always been an important research content in biomedical photonics.Nowadays,thermal treatment in clinical applications is the use of the effect between laser and biological tissue for diseased tissue coagulation.However,no matter whether the tissues are diseased or not,tissue damage will occur,when heating temperature greater than a certain range after a certain time.Optical Coherence Tomography(OCT)is a non-contact,non-invasive medical imaging technology with micron-level resolution,which can perform high-resolution imaging of internal structures of tissues.In recent years,artificial intelligence algorithms based on deep learning have continued to make breakthroughs,which has promoted the development of smart medicine rapidly.Constructing a segmentation algorithm based on deep learning can complete the rapid segmentation of the lesion location in the damaged tissue image,and it can also assist doctors in diagnosis and treatment.Therefore,using deep learning to learn the characteristics of tissue damage and segment the damage range of the damaged OCT image has an important application in assisting doctors' diagnosis and treatment.In this thesis,in order to cause different photothermal damage,a photothermal treatment system based on infrared thermal imager and PID control principle was designed.The system is mainly composed of infrared thermal imager,laser and Lab VIEW control module.Infrared thermal imager noninvasively measures the temperature of tissue damage range in real time The laser power is adjusted by PID algorithm to maintain the desired temperature.Finally,Lab VIEW was used to write a software program to control the instrument and virtualize the instrument panel and corresponding functions on the computer screen to display the temperature status of the measured tissue and laser power in real time.The images use optical coherence tomography(OCT)to obtain high-resolution images of damaged tissue.After acquiring the corresponding images,the diameter and depth of the photothermal damage of the tissue were recorded,which will be used to predict the size of the damage depth later.In order to complete the damage segmentation and depth prediction,we use a damage segmentation method based on a convolutional neural network to segment the external damage range of the tissue through training and learning of the data.At the same time,in order to make up for the inability to determine the boundaries of the internal damage,the regression model was used to predict the injury depth based on the diameter of the damage.Finally,the extent of the damage is determined through the segmentation and depth prediction of the external damage.