Converting an RGB Image to a Grayscale Image Using Neural Network
In this project, I have developed an neural network based autoencoder which learns the functions to convert a RGB image to Grayscale image. An autoencoder is a deep neural network which tries to copy its input to output. First encoder compresses the image to its latent space representation and then the decoder reconstructs the image from its latent space representation. I have used Undercomplete Autoencoder to achieve my desired goal. The autoencoder learns by minimizing the loss function. I have used Mean Squarred Error as my loss fuction. The model has been trained for 50 epochs.
