Fluorescent-assisted Fungus Automatic Microscopy System

Superficial fungal infections are extremely common in the diagnosis of clinical dermatological diseases.These fungi can be transmitted through direct contact or indirectly through epidermis and nails.According to relevant research,the rate of fungal infection in China increased from 13.9% in the early 1990 s to 24.4% in the late 1990 s.With the increasing number of patients with fungal infections,hospital dermatology technicians need to perform a large number of fungal detection or identification.So fast and accurate implementation of fungal detection is of great significance.Direct microscopy examination is the most frequently-used method for clinical diagnosis of superficial fungal infections due to the advantages including low cost,quickness and high positive rate.The basic method is to observe the prepared sample under a microscope: if the fungal structure(hyphae and spore)of the pathogenic phase is observed in the microscope,the fungal infection can be diagnosed.However,due to the low degree of automation of the equipment,the degree of manual intervention of direct microscopy examination is extremely high.When faced with a large number of samples to be examined,the workload of the technician will increase significantly,and the fatigue caused by such high-intensity work is likely to lead to misjudgment and missed judgment.In addition,the subjective degree of the microscopic examination and interpretation process is high,and there is no uniform evaluation standard.Different technicians may make different interpretation results for one sample,which may affect the final diagnosis of the attending doctor.Through the analysis of the operation process of direct microscopy examination,the shortcomings and the parts to be optimized are summarized,we finally proposes a scheme for automatic microscopy of fungi: the sample is prepared by using a fluorescent dye to specifically label the sample,at the same time constructing a fluorescence imaging structure with autonomous focusing and scanning imaging functions.Finally,a deep learning algorithm is adopted to analyze the collected microscopic image data of the fungus.We designed the fluorescent-assisted fungus automatic microscopy system by the combination of optical,mechanical,electronic and computer science concerned technique.Through the test of clinical samples,the detection time of typical positive samples in this system is about 3 to 5min,which is slightly longer than 2 to 3min of manual detection;but the accuracy of microscopic examination is quite similar to that of manual microscopy,which shows great potential for clinical application.