Skip to content

Using Caffe and python to reproduce the results of MTCNN on FDDB dataset.

Notifications You must be signed in to change notification settings

hualitlc/MTCNN-on-FDDB-Dataset

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

15 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

MTCNN-on-FDDB-Dataset

Using Caffe and python to reproduce the results of MTCNN on FDDB dataset.

References

  1. The implementation of MTCNN using python & caffe, thank the author DuinoDu/mtcnn.
  2. (alternatively) We can convert the ellipse annotations into the rectangle annotations for better visualization. Thank the author ankanbansal/fddb-for-yolo.

Better visualization

We can use the convertEllipseToRectangle.py to convert the FDDB-folds/FDDB-fold-01-ellipseList.txt into FDDB-folds/FDDB-fold-01-rectList.txt and show the converted bounding boxes on the images.

Run MTCNN on general image

We can use the demo.py to run mtcnn framwork on general images. This file comes from DuinoDu/mtcnn/demo.py.

Run MTCNN on FDDB dataset

We can use the runFDDB.py to run mtcnn framwork on FDDB dataset.

The official evaluation of your results

Download the official evaluation code and use the commond 'make' in the evaluation folder. To evaluate the results/preditions of your framework, just use the following codes.

./evaluate -a ../data/FDDB-folds/ellipseList.txt -d ../data/FDDB-folds/predict.txt -l ../data/FDDB-folds/foldList.txt -f 0

Then tempContROC.txt and tempDiscROC.txt will be generated in the /data/FDDB-folds/.

Draw the ROC curves

  1. Install the toolbox Gnuplot.
  2. Using the following commond.
gnuplot contROC.p 
gnuplot discROC.p

We will get the tempContROC-MTCNN.png(tempDiscROC-MTCNN.png) and the ROC curves like this: image image

Citation

@article{7553523,
    author={K. Zhang and Z. Zhang and Z. Li and Y. Qiao}, 
    journal={IEEE Signal Processing Letters}, 
    title={Joint Face Detection and Alignment Using Multitask Cascaded Convolutional Networks}, 
    year={2016}, 
    volume={23}, 
    number={10}, 
    pages={1499-1503}, 
    keywords={Benchmark testing;Computer architecture;Convolution;Detectors;Face;Face detection;Training;Cascaded convolutional neural network (CNN);face alignment;face detection}, 
    doi={10.1109/LSP.2016.2603342}, 
    ISSN={1070-9908}, 
    month={Oct}
}

About

Using Caffe and python to reproduce the results of MTCNN on FDDB dataset.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published