Journal of Biomedical Engineering and Medical Devices
Open Access

Department of Computer Science, Michael Okpara University of Agriculture Umudike, Umuahia, Abia State, Nigeria

Publications

• Research Article   
  The Effects of Modified ReLU Activation Functions in Image Classification
  Author(s): Charles Chinedu Nworu*, Emmanuel John Ekpenyong, John Chisimkwuo, Christian Nduka Onyeukwu, Godwin Okwara and Onyekachi Joy Agwu
  
  The choice of activation functions is very important in deep learning. This is because activation functions are capable of capturing non- linear patterns in a data. The most popular activation function is the Rectified Linear Unit (ReLU) but it suffers from gradient vanishing problem. Therefore, we examined the modifications of the ReLU activation function to determine its effectiveness (accuracy) and efficiency (time complexity). The effectiveness and efficiency was verified by conducting an empirical experiment using x-ray images that contains pneumonia and normal samples. Our experiments show that the modified ReLU, ReLU6 performed better in terms of low generalization error (97.05% training accuracy and 78.21% test accuracy). The sensitivity analysis also suggests that the ELU is capable of correctly predicting more than half of the positive cases with 52.14% .. View More»
  DOI: 10.35248/2475-7586.22.07.237

  Abstract HTML PDF