# DSOD: Learning Deeply Supervised Object Detectors from Scratch

arXiv: https://arxiv.org/abs/1708.01241 · HTML: https://mlsanity.com/paper/1708.01241/

## Key facts

- Cited by 41 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 3 Aug 2017.

## Built on by (most cited first, 30 of 41)

- [FAIR1M: A Benchmark Dataset for Fine-grained Object Recognition in High-Resolution Remote Sensing Imagery](https://mlsanity.com/paper/2103.05569/) (2021)
- [A Survey of Modern Deep Learning based Object Detection Models](https://mlsanity.com/paper/2104.11892/) (2021)
- [Cross Domain Object Detection by Target-Perceived Dual Branch Distillation](https://mlsanity.com/paper/2205.01291/) (2022)
- [Sliced Recursive Transformer](https://mlsanity.com/paper/2111.05297/) (2021)
- [Causal Reasoning Meets Visual Representation Learning: A Prospective Study](https://mlsanity.com/paper/2204.12037/) (2022)
- [Remote Sensing Object Detection Meets Deep Learning: A Meta-review of Challenges and Advances](https://mlsanity.com/paper/2309.06751/) (2023)
- [Circle Representation for Medical Object Detection](https://mlsanity.com/paper/2110.12093/) (2021)
- [Deep transfer learning for image classification: a survey](https://mlsanity.com/paper/2205.09904/) (2022)
- [AirDet: Few-Shot Detection without Fine-tuning for Autonomous Exploration](https://mlsanity.com/paper/2112.01740/) (2021)
- [Affine-Consistent Transformer for Multi-Class Cell Nuclei Detection](https://mlsanity.com/paper/2310.14154/) (2023)
- [A Comprehensive Review on Deep Supervision: Theories and Applications](https://mlsanity.com/paper/2207.02376/) (2022)
- [AirShot: Efficient Few-Shot Detection for Autonomous Exploration](https://mlsanity.com/paper/2404.05069/) (2024)
- [CenterNet++ for Object Detection](https://mlsanity.com/paper/2204.08394/) (2022)
- [Performance-aware Approximation of Global Channel Pruning for Multitask CNNs](https://mlsanity.com/paper/2303.11923/) (2023)
- [CE-FPN: Enhancing Channel Information for Object Detection](https://mlsanity.com/paper/2103.10643/) (2021)
- [DAP: Detection-Aware Pre-training with Weak Supervision](https://mlsanity.com/paper/2103.16651/) (2021)
- [Few-Shot Video Object Detection](https://mlsanity.com/paper/2104.14805/) (2021)
- [Joint Distribution Alignment via Adversarial Learning for Domain Adaptive Object Detection](https://mlsanity.com/paper/2109.09033/) (2021)
- [Deep Learning based Computer Vision Methods for Complex Traffic Environments Perception: A Review](https://mlsanity.com/paper/2211.05120/) (2022)
- [Radar Guided Dynamic Visual Attention for Resource-Efficient RGB Object Detection](https://mlsanity.com/paper/2206.01772/) (2022)
- [LEVI: Generalizable Fine-tuning via Layer-wise Ensemble of Different Views](https://mlsanity.com/paper/2402.04644/) (2024)
- [Revisiting the Loss Weight Adjustment in Object Detection](https://mlsanity.com/paper/2103.09488/) (2021)
- [Adversarial Vulnerability of Temporal Feature Networks for Object Detection](https://mlsanity.com/paper/2208.10773/) (2022)
- [Temporal Feature Networks for CNN based Object Detection](https://mlsanity.com/paper/2103.12213/) (2021)
- [Rethinking Training from Scratch for Object Detection](https://mlsanity.com/paper/2106.03112/) (2021)
- [Non-binary deep transfer learning for image classification](https://mlsanity.com/paper/2107.08585/) (2021)
- [Multitask Learning for SAR Ship Detection with Gaussian-Mask Joint Segmentation](https://mlsanity.com/paper/2411.13847/) (2024)
- [Delving into Variance Transmission and Normalization: Shift of Average Gradient Makes the Network Collapse](https://mlsanity.com/paper/2103.11590/) (2021)
- [Detecting Small Objects in Thermal Images Using Single-Shot Detector](https://mlsanity.com/paper/2108.11101/) (2021)
- [A Comprehensive Review of AI-enabled Unmanned Aerial Vehicle: Trends, Vision , and Challenges](https://mlsanity.com/paper/2310.16360/) (2023)

## Abstract (arXiv)

We present Deeply Supervised Object Detector (DSOD), a framework that can learn object detectors from scratch. State-of-the-art object objectors rely heavily on the off-the-shelf networks pre-trained on large-scale classification datasets like ImageNet, which incurs learning bias due to the difference on both the loss functions and the category distributions between classification and detection tasks. Model fine-tuning for the detection task could alleviate this bias to some extent but not fundamentally. Besides, transferring pre-trained models from classification to detection between discrepant domains is even more difficult (e.g. RGB to depth images). A better solution to tackle these two critical problems is to train object detectors from scratch, which motivates our proposed DSOD. Previous efforts in this direction mostly failed due to much more complicated loss functions and limited training data in object detection. In DSOD, we contribute a set of design principles for training object detectors from scratch. One of the key findings is that deep supervision, enabled by dense layer-wise connections, plays a critical role in learning a good detector. Combining with several other principles, we develop DSOD following the single-shot detection (SSD) framework. Experiments on PASCAL VOC 2007, 2012 and MS COCO datasets demonstrate that DSOD can achieve better results than the state-of-the-art solutions with much more compact models. For instance, DSOD outperforms SSD on all three benchmarks with real-time detection speed, while requires only 1/2 parameters to SSD and 1/10 parameters to Faster RCNN. Our code and models are available at: https://github.com/szq0214/DSOD .