Inicio » background subtraction
Archivos de la categoría: background subtraction
Dataset: Background Models Challenge 2012
Background Models Challenge 2012 dataset is intended for evaluating background subtraction algorithms. Its main characteristics are:
- The dataset is divided into:
- Set of videos for learning purposes with ground truth.
- Set of synthetic videos for evaluation purposes with encrypted ground truth.
- Set of real videos for evaluation purposes with encrypted ground truth.
- Software (not source code) to automatically test the background subtraction results (making use of the encrypted ground truth).
Code for background subtraction
- Background subtraction library. [Link 1] [Link 2]
- Based on mixture of Gaussians. Detect shadows.
- Publications:
- 2004 Improved adaptive Gausian mixture model for background subtraction.
- 2006 Efficient adaptive density estimation per image pixel for the task of background subtraction.
- Matlab y c/c++ code.