# Efficient Object Detection in Large Images using Deep Reinforcement Learning

arXiv: https://arxiv.org/abs/1912.03966 · HTML: https://mlsanity.com/paper/1912.03966/

## Key facts

- Cited by 21 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 9 Dec 2019.

## Built on by (most cited first, 21 of 21)

- [Object Detection in Aerial Images: A Large-Scale Benchmark and Challenges](https://mlsanity.com/paper/2102.12219/) (2021)
- [QueryDet: Cascaded Sparse Query for Accelerating High-Resolution Small Object Detection](https://mlsanity.com/paper/2103.09136/) (2021)
- [Rethinking Pseudo Labels for Semi-Supervised Object Detection](https://mlsanity.com/paper/2106.00168/) (2021)
- [IS-COUNT: Large-scale Object Counting from Satellite Images with Covariate-based Importance Sampling](https://mlsanity.com/paper/2112.09126/) (2021)
- [Sequential Voting with Relational Box Fields for Active Object Detection](https://mlsanity.com/paper/2110.11524/) (2021)
- [ObjectBox: From Centers to Boxes for Anchor-Free Object Detection](https://mlsanity.com/paper/2207.06985/) (2022)
- [GOHSP: A Unified Framework of Graph and Optimization-based Heterogeneous Structured Pruning for Vision Transformer](https://mlsanity.com/paper/2301.05345/) (2023)
- [A Visual Active Search Framework for Geospatial Exploration](https://mlsanity.com/paper/2211.15788/) (2022)
- [A Partially Supervised Reinforcement Learning Framework for Visual Active Search](https://mlsanity.com/paper/2310.09689/) (2023)
- [Video Violence Recognition and Localization Using a Semi-Supervised Hard Attention Model](https://mlsanity.com/paper/2202.02212/) (2022)
- [Fewer is More: Efficient Object Detection in Large Aerial Images](https://mlsanity.com/paper/2212.13136/) (2022)
- [RLSS: A Deep Reinforcement Learning Algorithm for Sequential Scene Generation](https://mlsanity.com/paper/2206.02544/) (2022)
- [FoSp: Focus and Separation Network for Early Smoke Segmentation](https://mlsanity.com/paper/2306.04474/) (2023)
- [Adaptive Data Transport Mechanism for UAV Surveillance Missions in Lossy Environments](https://mlsanity.com/paper/2410.10843/) (2024)
- [SaccadeCam: Adaptive Visual Attention for Monocular Depth Sensing](https://mlsanity.com/paper/2103.12981/) (2021)
- [You Better Look Twice: a new perspective for designing accurate detectors with reduced computations](https://mlsanity.com/paper/2107.10050/) (2021)
- [Building Coverage Estimation with Low-resolution Remote Sensing Imagery](https://mlsanity.com/paper/2301.01449/) (2023)
- [Integrating Saliency Ranking and Reinforcement Learning for Enhanced Object Detection](https://mlsanity.com/paper/2408.06803/) (2024)
- [Reinforcement Explanation Learning](https://mlsanity.com/paper/2111.13406/) (2021)
- [Saccade Mechanisms for Image Classification, Object Detection and Tracking](https://mlsanity.com/paper/2206.05102/) (2022)
- [Reinforcement Learning as a Parsimonious Alternative to Prediction Cascades: A Case Study on Image Segmentation](https://mlsanity.com/paper/2402.11760/) (2024)

## Abstract (arXiv)

Traditionally, an object detector is applied to every part of the scene of interest, and its accuracy and computational cost increases with higher resolution images. However, in some application domains such as remote sensing, purchasing high spatial resolution images is expensive. To reduce the large computational and monetary cost associated with using high spatial resolution images, we propose a reinforcement learning agent that adaptively selects the spatial resolution of each image that is provided to the detector. In particular, we train the agent in a dual reward setting to choose low spatial resolution images to be run through a coarse level detector when the image is dominated by large objects, and high spatial resolution images to be run through a fine level detector when it is dominated by small objects. This reduces the dependency on high spatial resolution images for building a robust detector and increases run-time efficiency. We perform experiments on the xView dataset, consisting of large images, where we increase run-time efficiency by 50% and use high resolution images only 30% of the time while maintaining similar accuracy as a detector that uses only high resolution images.