# Light-Head R-CNN: In Defense of Two-Stage Object Detector

arXiv: https://arxiv.org/abs/1711.07264 · HTML: https://mlsanity.com/paper/1711.07264/

## Key facts

- Cited by 30 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 20 Nov 2017.

## Built on by (most cited first, 30 of 30)

- [Object Detection in Aerial Images: A Large-Scale Benchmark and Challenges](https://mlsanity.com/paper/2102.12219/) (2021)
- [TPH-YOLOv5: Improved YOLOv5 Based on Transformer Prediction Head for Object Detection on Drone-captured Scenarios](https://mlsanity.com/paper/2108.11539/) (2021)
- [Deep Learning for UAV-based Object Detection and Tracking: A Survey](https://mlsanity.com/paper/2110.12638/) (2021)
- [LR-FPN: Enhancing Remote Sensing Object Detection with Location Refined Feature Pyramid Network](https://mlsanity.com/paper/2404.01614/) (2024)
- [A Survey on Physical Adversarial Attack in Computer Vision](https://mlsanity.com/paper/2209.14262/) (2022)
- [RingMo-Aerial: An Aerial Remote Sensing Foundation Model With Affine Transformation Contrastive Learning](https://mlsanity.com/paper/2409.13366/) (2024)
- [Prediction-Guided Distillation for Dense Object Detection](https://mlsanity.com/paper/2203.05469/) (2022)
- [Computation-efficient Deep Learning for Computer Vision: A Survey](https://mlsanity.com/paper/2308.13998/) (2023)
- [A Comprehensive Study of Real-Time Object Detection Networks Across Multiple Domains: A Survey](https://mlsanity.com/paper/2208.10895/) (2022)
- [Dynamic Sparse R-CNN](https://mlsanity.com/paper/2205.02101/) (2022)
- [PrObeD: Proactive Object Detection Wrapper](https://mlsanity.com/paper/2310.18788/) (2023)
- [Deep Learning for Weakly-Supervised Object Detection and Object Localization: A Survey](https://mlsanity.com/paper/2105.12694/) (2021)
- [Scale Normalized Image Pyramids with AutoFocus for Object Detection](https://mlsanity.com/paper/2102.05646/) (2021)
- [Decoupled IoU Regression for Object Detection](https://mlsanity.com/paper/2202.00866/) (2022)
- [R(Det)^2: Randomized Decision Routing for Object Detection](https://mlsanity.com/paper/2204.00794/) (2022)
- [Revisiting Knowledge Distillation for Object Detection](https://mlsanity.com/paper/2105.10633/) (2021)
- [LO-Det: Lightweight Oriented Object Detection in Remote Sensing Images](https://mlsanity.com/paper/2209.07709/) (2022)
- [TinyDet: Accurate Small Object Detection in Lightweight Generic Detectors](https://mlsanity.com/paper/2304.03428/) (2023)
- [A More Compact Object Detector Head Network with Feature Enhancement and Relational Reasoning](https://mlsanity.com/paper/2106.14475/) (2021)
- [Deep Learning Computer Vision Algorithms for Real-time UAVs On-board Camera Image Processing](https://mlsanity.com/paper/2211.01037/) (2022)
- [Image-Based Fire Detection in Industrial Environments with YOLOv4](https://mlsanity.com/paper/2212.04786/) (2022)
- [LGA-RCNN: Loss-Guided Attention for Object Detection](https://mlsanity.com/paper/2104.13763/) (2021)
- [Identification of Driver Phone Usage Violations via State-of-the-Art Object Detection with Tracking](https://mlsanity.com/paper/2109.02119/) (2021)
- [Selective Multi-Scale Learning for Object Detection](https://mlsanity.com/paper/2206.08206/) (2022)
- [Delving into the Scale Variance Problem in Object Detection](https://mlsanity.com/paper/2206.08227/) (2022)
- [HPS-Det: Dynamic Sample Assignment with Hyper-Parameter Search for Object Detection](https://mlsanity.com/paper/2207.11539/) (2022)
- [Detecting Objects with Context-Likelihood Graphs and Graph Refinement](https://mlsanity.com/paper/2212.12395/) (2022)
- [Evaluation of Environmental Conditions on Object Detection using Oriented Bounding Boxes for AR Applications](https://mlsanity.com/paper/2306.16798/) (2023)
- [A Study on Real-time Object Detection using Deep Learning](https://mlsanity.com/paper/2602.15926/) (2026)
- [DTKDP: A Dual Teacher Knowledge Distillation and Pruning Framework for Lightweight Oriented SAR Ship Detection](https://mlsanity.com/paper/2609.24872/) (2026)

## Abstract (arXiv)

In this paper, we first investigate why typical two-stage methods are not as fast as single-stage, fast detectors like YOLO and SSD. We find that Faster R-CNN and R-FCN perform an intensive computation after or before RoI warping. Faster R-CNN involves two fully connected layers for RoI recognition, while R-FCN produces a large score maps. Thus, the speed of these networks is slow due to the heavy-head design in the architecture. Even if we significantly reduce the base model, the computation cost cannot be largely decreased accordingly. We propose a new two-stage detector, Light-Head R-CNN, to address the shortcoming in current two-stage approaches. In our design, we make the head of network as light as possible, by using a thin feature map and a cheap R-CNN subnet (pooling and single fully-connected layer). Our ResNet-101 based light-head R-CNN outperforms state-of-art object detectors on COCO while keeping time efficiency. More importantly, simply replacing the backbone with a tiny network (e.g, Xception), our Light-Head R-CNN gets 30.7 mmAP at 102 FPS on COCO, significantly outperforming the single-stage, fast detectors like YOLO and SSD on both speed and accuracy. Code will be made publicly available.