Skip to content

Latest commit

 

History

History
62 lines (37 loc) · 2.58 KB

README.md

File metadata and controls

62 lines (37 loc) · 2.58 KB

DSSD MXNET

Features

Specific VOC mAP Results

backbone SSD SSD_min_loss DSSD* DSSD_stage1 DSSD_stage2 SSD+TDM
vgg16-512 75.56 75.35 75.85 65.16 75.77 76.80
vgg16-300 74.65 74.59 75.74 —— —— ——
resnet101-512 78.43 78.02 79.25 71.98 78.19 79.18
resnet101-321 75.18 74.80 75.54 —— —— ——
  • DSSD*: results by use our traning strategy,for more details,please see here

Requirements

We tested our code on:

Ubuntu 16.04, Python 2.7 with

numpy(1.11.0), cv2(3.3.0-dev)

mxnet 0.11.0

Preparation for Training

1.Download the converted pretrained vgg16_reduced model here.

2.Prepare VOC datasets and generate .rec files by using tools/prepare_pascal.sh

3.Set TDM or DSSD mode in function get_config from symbol/symbol_factory.py.By default, DSSD mode is used,please set all configs by your needs.

4.start training

python train.py

or choice a bash file which provide in ./script/ to run some default parameters setting,such like try the two stage training strategy.

bash scripts/stage1_dssd_train_res_voc.sh

Demo

  1. Download model, available at here, and place it in the model folder.
  2. run demo.py

References

1.SSD: Liu W, Anguelov D, Erhan D, Szegedy C, Reed S, Fu CY, Berg AC. SSD: Single shot multibox detector. InEuropean Conference on Computer Vision 2016 Oct 8 (pp. 21-37). Springer International Publishing.Link

2.DSSD: Fu CY, Liu W, Ranga A, Tyagi A, Berg AC. DSSD: Deconvolutional Single Shot Detector. arXiv preprint arXiv:1701.06659. 2017 Jan 23. Link

3.TDM: Shrivastava A, Sukthankar R, Malik J, Gupta A. Beyond Skip Connections: Top-Down Modulation for ObjectDetection. arXiv preprint arXiv:1612.06851. 2016 Dec 20.Link