# A Simple Approach to Image Tilt Correction with Self-Attention MobileNet for Smartphones

arXiv: https://arxiv.org/abs/2111.00398 · HTML: https://mlsanity.com/paper/2111.00398/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 23 of its 36 references that are in the map.
- First posted on arXiv 31 Oct 2021.

## Builds on (most cited first, 23 of 23)

- [Attention Is All You Need](https://mlsanity.com/paper/1706.03762/) (2017)
- [Deep Residual Learning for Image Recognition](https://mlsanity.com/paper/1512.03385/) (2015)
- [Very Deep Convolutional Networks for Large-Scale Image Recognition](https://mlsanity.com/paper/1409.1556/) (2014)
- [End-to-End Object Detection with Transformers](https://mlsanity.com/paper/2005.12872/) (2020)
- [Rethinking the Inception Architecture for Computer Vision](https://mlsanity.com/paper/1512.00567/) (2015)
- [MobileNetV2: Inverted Residuals and Linear Bottlenecks](https://mlsanity.com/paper/1801.04381/) (2018)
- [Squeeze-and-Excitation Networks](https://mlsanity.com/paper/1709.01507/) (2017)
- [CBAM: Convolutional Block Attention Module](https://mlsanity.com/paper/1807.06521/) (2018)
- [Searching for MobileNetV3](https://mlsanity.com/paper/1905.02244/) (2019)
- [Show, Attend and Tell: Neural Image Caption Generation with Visual Attention](https://mlsanity.com/paper/1502.03044/) (2015)
- [Self-Attention Generative Adversarial Networks](https://mlsanity.com/paper/1805.08318/) (2018)
- [Stacked Attention Networks for Image Question Answering](https://mlsanity.com/paper/1511.02274/) (2015)
- [A Decomposable Attention Model for Natural Language Inference](https://mlsanity.com/paper/1606.01933/) (2016)
- [BAM: Bottleneck Attention Module](https://mlsanity.com/paper/1807.06514/) (2018)
- [Long Short-Term Memory-Networks for Machine Reading](https://mlsanity.com/paper/1601.06733/) (2016)
- [A Comprehensive Analysis of Deep Regression](https://mlsanity.com/paper/1803.08450/) (2018)
- [A Perceptual Measure for Deep Single Image Camera Calibration](https://mlsanity.com/paper/1712.01259/) (2017)
- [Surface Normal Estimation of Tilted Images via Spatial Rectifier](https://mlsanity.com/paper/2007.09264/) (2020)
- [Horizon Lines in the Wild](https://mlsanity.com/paper/1604.02129/) (2016)
- [Detecting Vanishing Points using Global Image Context in a Non-Manhattan World](https://mlsanity.com/paper/1608.05684/) (2016)
- [UprightNet: Geometry-Aware Camera Orientation Estimation from Single Images](https://mlsanity.com/paper/1908.07070/) (2019)
- [Deep Image Orientation Angle Detection](https://mlsanity.com/paper/2007.06709/) (2020)
- [ULSAM: Ultra-Lightweight Subspace Attention Module for Compact Convolutional Neural Networks](https://mlsanity.com/paper/2006.15102/) (2020)

## Built on by (most cited first, 1 of 1)

- [Deep Learning for Camera Calibration and Beyond: A Survey](https://mlsanity.com/paper/2303.10559/) (2023)

## Abstract (arXiv)

The main contributions of our work are two-fold. First, we present a Self-Attention MobileNet, called SA-MobileNet Network that can model long-range dependencies between the image features instead of processing the local region as done by standard convolutional kernels. SA-MobileNet contains self-attention modules integrated with the inverted bottleneck blocks of the MobileNetV3 model which results in modeling of both channel-wise attention and spatial attention of the image features and at the same time introduce a novel self-attention architecture for low-resource devices. Secondly, we propose a novel training pipeline for the task of image tilt detection. We treat this problem in a multi-label scenario where we predict multiple angles for a tilted input image in a narrow interval of range 1-2 degrees, depending on the dataset used. This process induces an implicit correlation between labels without any computational overhead of the second or higher-order methods in multi-label learning. With the combination of our novel approach and the architecture, we present state-of-the-art results on detecting the image tilt angle on mobile devices as compared to the MobileNetV3 model. Finally, we establish that SA-MobileNet is more accurate than MobileNetV3 on SUN397, NYU-V1, and ADE20K datasets by 6.42%, 10.51%, and 9.09% points respectively, and faster by at least 4 milliseconds on Snapdragon 750 Octa-core.