Mobile Nets: Prediction of Black Fungus Disease Through Image Classification
摘要
Mucormycosis, an uncommon fungal illness, may be lethal. If it is not recognized and treated swiftly in the early stages, between 50 and 80% of individuals might die from it. This condition typically arises in people whose immune systems have been compromised as a result of exposure to the coronavirus (COVID-19), other viral diseases, immunodeficiency disorders, cancers, chronic diseases, or other medical conditions, or as a result of the use of anti-cancer medications or other medications that suppress the immune system. MobileNet v2 architecture is a transfer learning method adopted to recognize black fungus infections based on photographs. It is less computationally costly and an excellent match for mobile devices and computers without GPU deployment. When compared to more complicated learning-based approaches, the suggested filter significance may be used to search the per-layer pruning ratio, displaying results that are either equivalent to or superior to those achieved after fine-tuning.