# Improving the Energy Efficiency and Robustness of tinyML Computer Vision using Log-Gradient Input Images

arXiv: https://arxiv.org/abs/2203.02571 · HTML: https://mlsanity.com/paper/2203.02571/

## Key facts

- Cited by 2 papers in the map (data through 14 Oct 2026).
- Builds on 10 of its 27 references that are in the map.
- First posted on arXiv 4 Mar 2022.

## Builds on (most cited first, 10 of 10)

- [MCUNet: Tiny Deep Learning on IoT Devices](https://mlsanity.com/paper/2007.10319/) (2020)
- [MLPerf Tiny Benchmark](https://mlsanity.com/paper/2106.07597/) (2021)
- [ExpandNet: A Deep Convolutional Neural Network for High Dynamic Range Expansion from Low Dynamic Range Content](https://mlsanity.com/paper/1803.02266/) (2018)
- [μNAS: Constrained Neural Architecture Search for Microcontrollers](https://mlsanity.com/paper/2010.14246/) (2020)
- [Visual Wake Words Dataset](https://mlsanity.com/paper/1906.05721/) (2019)
- [Reconfiguring the Imaging Pipeline for Computer Vision](https://mlsanity.com/paper/1705.04352/) (2017)
- [OpenLORIS-Object: A Robotic Vision Dataset and Benchmark for Lifelong Deep Learning](https://mlsanity.com/paper/1911.06487/) (2019)
- [RNNPool: Efficient Non-linear Pooling for RAM Constrained Inference](https://mlsanity.com/paper/2002.11921/) (2020)
- [Euphrates: Algorithm-SoC Co-Design for Low-Power Mobile Continuous Vision](https://mlsanity.com/paper/1803.11232/) (2018)
- [FQ-Conv: Fully Quantized Convolution for Efficient and Accurate Inference](https://mlsanity.com/paper/1912.09356/) (2019)

## Built on by (most cited first, 2 of 2)

- [From Tiny Machine Learning to Tiny Deep Learning: A Survey](https://mlsanity.com/paper/2506.18927/) (2025)
- [TinyReptile: TinyML with Federated Meta-Learning](https://mlsanity.com/paper/2304.05201/) (2023)

## Abstract (arXiv)

This paper studies the merits of applying log-gradient input images to convolutional neural networks (CNNs) for tinyML computer vision (CV). We show that log gradients enable: (i) aggressive 1.5-bit quantization of first-layer inputs, (ii) potential CNN resource reductions, and (iii) inherent robustness to illumination changes (1.7% accuracy loss across 1/32...8 brightness variation vs. up to 10% for JPEG). We establish these results using the PASCAL RAW image data set and through a combination of experiments using neural architecture search and a fixed three-layer network. The latter reveal that training on log-gradient images leads to higher filter similarity, making the CNN more prunable. The combined benefits of aggressive first-layer quantization, CNN resource reductions, and operation without tight exposure control and image signal processing (ISP) are helpful for pushing tinyML CV toward its ultimate efficiency limits.