WaveMixSR: A Resource-efficient Neural Network for Image Super-resolution
This work addresses resource efficiency for image super-resolution tasks, offering a more practical alternative to transformer models, though it is incremental as it builds on existing WaveMix architecture.
The authors tackled the problem of high computational resource requirements in transformer-based image super-resolution models by proposing WaveMixSR, a neural network that uses 2D-discrete wavelet transform for spatial token-mixing, achieving competitive performance with less training data and resources, including state-of-the-art results on the BSD100 dataset.
Image super-resolution research recently been dominated by transformer models which need higher computational resources than CNNs due to the quadratic complexity of self-attention. We propose a new neural network -- WaveMixSR -- for image super-resolution based on WaveMix architecture which uses a 2D-discrete wavelet transform for spatial token-mixing. Unlike transformer-based models, WaveMixSR does not unroll the image as a sequence of pixels/patches. It uses the inductive bias of convolutions along with the lossless token-mixing property of wavelet transform to achieve higher performance while requiring fewer resources and training data. We compare the performance of our network with other state-of-the-art methods for image super-resolution. Our experiments show that WaveMixSR achieves competitive performance in all datasets and reaches state-of-the-art performance in the BSD100 dataset on multiple super-resolution tasks. Our model is able to achieve this performance using less training data and computational resources while maintaining high parameter efficiency compared to current state-of-the-art models.