Black Fungus Infection Detection using AI-based Early Warning System for Patients through Multi-Modal Medical Imaging

Abstract

The COVID-19 pandemic has inflicted widespread devastation worldwide, causing a significant loss of human lives and impacting economies. Amid the ongoing pandemic, a perilous fungal infection known as mucormycosis, or “black fungus,” emerged as a serious concern during the second wave of COVID-19 in 2021. This disease primarily affects individuals already battling other illnesses and undergoing intensive medication, leading to a weakened immune response against fungal infections. Mucormycosis predominantly manifests on the skin but can also affect vital organs such as the eyes and brain. This research paper presents a novel approach to detect and diagnose mucormycosis using artificial intelligence (AI) techniques, specifically a modified neural network logic integrating Regional Convolutional Neural Networks (RCNN) and Support Vector Machine (SVM). The study utilizes a dataset containing a collection of eye photographs obtained from COVID-19 patients, some of whom developed mucormycosis during their illness. The research methodology encompasses key steps, including image acquisition, preprocessing, feature extraction, and classification, all aligned with established principles of dataset training and testing. The results of this approach are presented graphically, providing precise specifications, and highlighting the effectiveness of the proposed AI-based method. Signs and symptoms of mucormycosis are detailed, emphasizing the importance of early diagnosis and specialized tests to combat this potentially life-threatening infection. The clinical pathogenesis of mucormycosis, its risk factors, and treatment strategies are explored in-depth. Additionally, the paper underscores the increased susceptibility to mucormycosis among COVID-19 patients with certain comorbidities. Furthermore, the study discusses the challenges in diagnosing mucormycosis and highlights the significance of advanced techniques such as computed tomography (CT) scans, magnetic resonance imaging (MRI) scans, and cell biopsy tests. Image acquisition and preprocessing are elaborated upon to provide insights into the data preparation process, followed by feature extraction and classification techniques used to identify mucormycosis in images. The dataset utilized in this research is described, including its source and composition, and the performance of the AI-based method is evaluated in terms of accuracy, sensitivity, precision, and recall. Notably, the approach achieves an overall accuracy of 81.65%, demonstrating its potential as a valuable tool in the early detection of mucormycosis. In conclusion, this research paper sheds light on the critical issue of mucormycosis in COVID-19 patients and presents an innovative AI-based solution for its detection and diagnosis. The findings emphasize the importance of timely intervention and highlight the potential of AI in enhancing healthcare during pandemics.

Topics & Concepts