Study Of Image Detection Methods And Devices In Portable In Vitro Diagnostics

Disease poses a serious threat to human life and health,and disease diagnosis is a prerequisite for disease treatment,prognosis,and prevention,and is of paramount importance.Traditional diagnostic examinations are carried out in various hospitals or medical centers using large scale testing equipment and by professional examiners,which often takes longer and costs more and is not conducive to the timely treatment of infectious diseases.In recent years,Point-of-Care Testing(POCT)has become increasingly popular and used as a new testing method.Microfluidics-based point-of-care testing methods have significant advantages in terms of reducing the number of reagents used,simplifying the testing process,reducing testing time and reducing testing costs.Image detection,as a typical signal acquisition and processing method,is playing an increasingly important role in biomedical testing,especially in the field of in vitro diagnostics.The improved image detection method,applied to the field rapid inspection device,is of great value in improving its overall detection performance and level.An image processing algorithm for lateral flow immunocolloidal gold chromatography strips was developed and implemented for HIV detection based on lateral flow immunocolloidal gold chromatography strips.Research into the implementation of optimization algorithms based on traditional image processing methods.Based on this,further research was conducted to implement an improved algorithm based on deep learning,and a distributed HIV detection system was constructed to achieve local acquisition and remote processing and analysis of paper strip images.A hybrid algorithm model has been built based on deep learning algorithms,combining artificial intelligence models with traditional image recognition algorithms,significantly improving the detection accuracy of low-positive test strips,and all types of special negative or positive test strips.Five tests for eugenics and fertility(TORCH test),namely point-of-care testing of five pathogens including TOX,RV,CMV,HSV-I and HSV-2.We have researched and implemented a chemiluminescence image detection and processing algorithm for reaction microspheres,and implemented algorithm software design.The chemiluminescence image processing algorithm mainly includes functional modules such as image preprocessing algorithm,circle recognition algorithm,and circle correction algorithm.By optimizing the algorithm,the impact of factors such as fluid mixing,bubbles,and fluctuations in microsphere position on image detection has been effectively overcome,significantly improving the accuracy of microsphere chemiluminescence signal detection.A paper based microfluidic chip and related fluorescence image processing algorithms were studied for isothermal nucleic acid amplification.A two-step isothermal amplification paper based microfluidic chip based on RPA amplification was studied,which can be used for multi-indicator joint detection.By combining the liquid amplification chamber with the paperbased amplification matrix,the fluid driven two-step and multi-indicator detection of amplification reactions is effectively simplified.At the same time,using a two-step reaction mode is beneficial for improving detection sensitivity.A portable detection device and fluorescence image processing algorithm have been developed for paper based isothermal amplification chips,enabling rapid collection and analysis of fluorescence signals on paperbased reaction carriers.In principle,it has been verified that based on the paper based isothermal amplification chip,two-step and multi-indicator detection can be achieved.