Scholarlink

Malaria is a significant global health concern affecting millions of people annually, especially in resource-limited areas. For effective treatment, rapid and accurate detection of malaria through blood smear analysis is required. Traditional microscopic-based malaria diagnosis is time-consuming and relies on dependence on trained specialists, making it challenging for many healthcare facilities in limited resource areas. This research work proposed a novel Custom Interpretable Malaria Diagnostic (CIMD) model that integrates a convolutional neural network (CNN) for extracting the features and support vector machine (SVM) for the classification of infected and uninfected malarial cell images. The proposed CIMD is evaluated with other classifiers such as CNN + K-NN and CNN + DT with various metrics such as accuracy, precision, recall, F-score, false positive rate (FPR), false negative rate (FNR), and diagnostic odds ratio (DOR). The results revealed that the proposed method achieved an accuracy of 95.28% and precision of 95.73% demonstrating its efficiency in detecting malaria cells. Additionally, Gradient-Weighted Class Activation Mapping is used to interpret the features selected by the proposed CIMD.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Custom Interpretable Malaria Diagnosis Using CNN and SVM with Grad-CAM Visualization

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Mohammed Ismail Abdul Azeez,
  • Rajalakshmi Shenbaga Moorthy

摘要

Malaria is a significant global health concern affecting millions of people annually, especially in resource-limited areas. For effective treatment, rapid and accurate detection of malaria through blood smear analysis is required. Traditional microscopic-based malaria diagnosis is time-consuming and relies on dependence on trained specialists, making it challenging for many healthcare facilities in limited resource areas. This research work proposed a novel Custom Interpretable Malaria Diagnostic (CIMD) model that integrates a convolutional neural network (CNN) for extracting the features and support vector machine (SVM) for the classification of infected and uninfected malarial cell images. The proposed CIMD is evaluated with other classifiers such as CNN + K-NN and CNN + DT with various metrics such as accuracy, precision, recall, F-score, false positive rate (FPR), false negative rate (FNR), and diagnostic odds ratio (DOR). The results revealed that the proposed method achieved an accuracy of 95.28% and precision of 95.73% demonstrating its efficiency in detecting malaria cells. Additionally, Gradient-Weighted Class Activation Mapping is used to interpret the features selected by the proposed CIMD.