Image Denoising with Autoencoders in R (University Project) Built a convolutional autoencoder in R using Keras/TensorFlow to perform image denoising on MNIST and CIFAR-10 datasets with varying levels of Gaussian noise. Achieved up to 33.6 dB PSNR on MNIST (σ = 0.1). Using PSNR, SSIM to evaluate the model. Technologies: R, Keras, TensorFlow
-
Updated
Jul 30, 2025 - R