A Novel Self Learning Activation Function Using Gegenbauer Polynomials for An Effective Medical Image Denoising

In medical image analysis denoising of an image is a salient preprocessing step. A large number of methods are proposed for image denoising with varying performance. In this research article, we present an innovative sparse-aware autoencoder model designed for effectively denoising images in non-Gaussian noise environments. The proposed model uses a self learning Gegenbauer polynomials based activation function. The performance of proposed model is verified on X-ray and CT scan images under various noise levels and observed its superiority over existing methods.