Документ взят из кэша поисковой машины. Адрес оригинального документа : http://graphics.cs.msu.ru/ru/science/research/machinelearning/dbcreator
Дата изменения: Sun Apr 10 02:17:49 2016
Дата индексирования: Sun Apr 10 02:17:49 2016
Кодировка: UTF-8
GML OpenCV Training DBCreator | Лаборатория компьютерной графики и мультимедиа

GML OpenCV Training DBCreator

Download

Download GML OpenCV Training DBCreator 1.21b

Introduction

GML DBCreator is a FREE software that enables to create an image base for training a facial features detector based on the Viola-Jones algorithm. Using this tool you can manually mark facial features out in individual images or video frames.

Nowadays facial features detection can be treated either as an independent significant problem or as an important stage in many complicated tasks in the field of computer vision. During our researches we examined several approaches to the problem and finally adopted the algorithm introduced by Viola and Jones [1]. It is an efficient object detection method based on a boosted cascade of weak classifiers.

This algorithm requires to be trained with a set of facial features images. For reliable detection the train set should be composed of several thousand pictures. Consequently, there is a challenge to make up such a base with minimal costs.

In order to simplify this task we developed GML DBCreator. This application makes it easy to create a high-quality training image set from video files. For further algorithm training the image base is saved in the special data format, which corresponds to the open source Intel OpenCV [2] realization of the Viola-Jones detector.

 

The project team

  • Rostislav Shorgin
  • Alexander Vezhnevets

Contacts

Please mail all comments, suggestions, problems and contributions:
vision@graphics.cs.msu.ru

References

[1] P. Viola and M. Jones. “Robust Real-time Object Detection.” In Proc. 2nd Int'l Workshop on Statistical and Computational Theories of Vision - Modeling, Learning, Computing and Sampling, Vancouver, Canada, July 2001.
[2] http://www.intel.com/technology/computing/opencv/index.htm (Open Computer Vision Library)