Review — DSOD: Learning Deeply Supervised Object Detectors from Scratch (Object Detection)

1st Detection Network Trained From Scratch, Modified From SSD, Using Dense Blocks From DenseNet, Outperforms SSD, YOLOv2, Faster R-CNN, R-FCN, ION

In this story, DSOD: Learning Deeply Supervised Object Detectors from Scratch, (DSOD), by Fudan University, Tsinghua University, and Intel Labs China, is reviewed. In this paper:

This is the paper in 2017 ICCV with over 300 citations. (Sik-Ho Tsang @ Medium)


1. DSOD: Network Architecture

DSOD prediction layers with plain and dense structures (300×300 input).

1.1. Backbone Subnetwork

1.2. Front-End Sub-Network

DSOD architecture (growth rate k = 48 in each dense block).

2. A Set of Principles to Train from Scratch

2.1. Principle 1: Proposal-Free

2.2. Principle 2: Deep Supervision

2.3. Principle 3: Stem Block

2.4. Principle 4: Dense Prediction Structure

3. Ablation Study

Ablation study on PASCAL VOC 2007 test set.
Details of Ablation study on PASCAL VOC 2007 test set. (DS/A-B-k-θ: Backbone network)

3.1. Configurations in Dense Blocks

3.2. Effectiveness of Design Principles

3.3. Runtime Analysis

4. Experimental Results

4.1. PASCAL VOC2007

PASCAL VOC 2007 test detection results.

4.2. PASCAL VOC2012

PASCAL VOC 2012 test detection results.

4.3. MS COCO

MS COCO test-dev 2015 detection results.

5. Discussions

5.1. Better Model Structure vs. More Training Data

Particularly, DSOD is only trained with 16,551 images on VOC 2007, but achieves competitive or even better performance than those models trained with 1.2 million + 16,551 images.

5.2. Why Training from Scratch?

First, there may be big domain differences from pre-trained model domain to the target one.

Second, model fine-tuning restricts the structure design space for object detection networks.

5.3. Model Compactness vs. Performance

For instance, the smallest dense model (DS/64–64–16–1, with dense prediction layers) achieves 73.6% mAP with only 5.9M parameters, which shows great potential for applications on low-end devices.

PhD, Researcher. I share what I've learnt and done. :) My LinkedIn:, My Paper Reading List:

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store