# High Frequency Residual Learning for Multi-Scale Image Classification

arXiv: https://arxiv.org/abs/1905.02649 · HTML: https://mlsanity.com/paper/1905.02649/

## Key facts

- Cited by 4 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 7 May 2019.

## Built on by (most cited first, 4 of 4)

- [Pro-UIGAN: Progressive Face Hallucination from Occluded Thumbnails](https://mlsanity.com/paper/2108.00602/) (2021)
- [Hybrid Skip: A Biologically Inspired Skip Connection for the UNet Architecture](https://mlsanity.com/paper/2207.04721/) (2022)
- [Multiscale Structure-Guided Latent Diffusion for Multimodal MRI Translation](https://mlsanity.com/paper/2603.12581/) (2026)
- [Single Cell Training on Architecture Search for Image Denoising](https://mlsanity.com/paper/2212.06368/) (2022)

## Abstract (arXiv)

We present a novel high frequency residual learning framework, which leads to a highly efficient multi-scale network (MSNet) architecture for mobile and embedded vision problems. The architecture utilizes two networks: a low resolution network to efficiently approximate low frequency components and a high resolution network to learn high frequency residuals by reusing the upsampled low resolution features. With a classifier calibration module, MSNet can dynamically allocate computation resources during inference to achieve a better speed and accuracy trade-off. We evaluate our methods on the challenging ImageNet-1k dataset and observe consistent improvements over different base networks. On ResNet-18 and MobileNet with alpha=1.0, MSNet gains 1.5% accuracy over both architectures without increasing computations. On the more efficient MobileNet with alpha=0.25, our method gains 3.8% accuracy with the same amount of computations.