Exploiting the Potential of Standard Convolutional Autoencoders for Image Restoration by Evolutionary Search
arXiv:1803.00370
Abstract
Researchers have applied deep neural networks to image restoration tasks, in which they proposed various network architectures, loss functions, and training methods. In particular, adversarial training, which is employed in recent studies, seems to be a key ingredient to success. In this paper, we show that simple convolutional autoencoders (CAEs) built upon only standard network components, i.e., convolutional layers and skip connections, can outperform the state-of-the-art methods which employ adversarial training and sophisticated loss functions. The secret is to employ an evolutionary algorithm to automatically search for good architectures. Training optimized CAEs by minimizing the loss between reconstructed images and their ground truths using the ADAM optimizer is all we need. Our experimental results show that this approach achieves 27.8 dB peak signal to noise ratio (PSNR) on the CelebA dataset and 40.4 dB on the SVHN dataset, compared to 22.8 dB and 33.0 dB provided by the former state-of-the-art methods, respectively.
arXiv commentOur code is available at https://github.com/sg-nm/Evolutionary-Autoencoders
References in corpus (10)
- Beyond a Gaussian Denoiser: Residual Learning of Deep CNN for Image Denoising
- Neural Architecture Search with Reinforcement Learning
- Depth Map Prediction from a Single Image using a Multi-Scale Deep Network
- Image Restoration Using Very Deep Convolutional Encoder-Decoder Networks with Symmetric Skip Connections
- Training Very Deep Networks
- MemNet: A Persistent Memory Network for Image Restoration
- Progressive Neural Architecture Search
- Semantic Image Inpainting with Deep Generative Models
- A Genetic Programming Approach to Designing Convolutional Neural Network Architectures
- Multi-Scale Continuous CRFs as Sequential Deep Networks for Monocular Depth Estimation