Background Subtraction

1. Background subtraction presents the following steps:

• • Background modeling

  • Background initialization

  • Background maintenance

  • Foreground detection

2. Background subtraction presents the following issues (Features Website):

• • Choice of the feature size: pixel, a block or a cluster.

  • Choice of the feature type: color features, edge features, stereo features, motion features and texture features.

3. Background subtraction presents solved an unsolved challenges. Please see Challenges in Background Subtraction Website (257 papers).

4. Background subtraction employs the following modeling concepts:

• • Mathematical concepts:  Crisp concepts, Statistical concepts, Fuzzy Concepts (Zadeh 1965), Dempster-Schafer (Dempster 1968,  Schafer 1976). The last three concepts allows to deal with imprecision, uncertainty and incompletness in the data.

  • Machine Learning Concepts: 

    • Representation Learning (Dimensionality Reduction, Subspace Learning): Conventional Subspace Learning (Pearson 1901), Robust Subspace Learning (Candès et al. 2009),  Dynamic Subspace Learning (Vaswani et al. 2010). Data are considered as inliers data, outliers data or missing data.

    • Neural Networks Learning:  Conventional Neural Networks (Rosenblatt 1957) , Support Vector Machines (Vapnik et al. 1995), Deep Neural Networks (Le Cun). Date are viewed as learning entities that can be well labeled or noisy labeled.

  • Signal Processing Concepts: Estimation Filters (Kalman 1960),  Sparse Representation, Graph Signal Processing (Ortega et al. 2017). Data are viewed as signal data (1D, 2D, or 3D).

  • Classification/Clustering Concepts: Crisp concepts, Statistical concepts, Fuzzy concepts (Bezdek 1978). Data are viewed as belonging or not to  a class or a cluster.