Growing a multi-class classifier with a reject option

• Published in: Pattern recognition letters Vol. 29; no. 10; pp. 1565 - 1570
• Main Authors: Tax, D.M.J., Duin, R.P.W.
• Format: Journal Article
• Language: English
• Published: Amsterdam Elsevier B.V 15.07.2008; Elsevier
• Subjects: Applied sciences; Exact sciences and technology; Image processing; Information, signal and communications theory; Multi-class classification; Outlier detection; Pattern recognition; Rejection; Signal and communications theory; Signal processing; Signal representation. Spectral analysis; Signal, noise; Telecommunications and information theory; Rejection; Multi-class classification; Outlier detection; Biometrics; Performance evaluation; Automatic classification; Face recognition; Image processing; Multi-class classification;Outlier detection;Rejection; Pattern recognition; Signal classification; Non linear transformation; Outlier; Probability density; Posterior probability; Heuristic method; Automatic recognition
• ISSN: 0167-8655; 1872-7344
• DOI: 10.1016/j.patrec.2008.03.010

In many classification problems objects should be rejected when the confidence in their classification is too low. An example is a face recognition problem where the faces of a selected group of people have to be classified, but where all other faces and non-faces should be rejected. These problems are typically solved by estimating the class densities and assigning an object to the class with the highest posterior probability. The total probability density is thresholded to detect the outliers. Unfortunately, this procedure does not easily allow for class-dependent thresholds, or for class models that are not based on probability densities but on distances. In this paper we propose a new heuristic to combine any type of one-class models for solving the multi-class classification problem with outlier rejection. It normalizes the average model output per class, instead of the more common non-linear transformation of the distances. It creates the possibility to adjust the rejection threshold per class, and also to combine class models that are not (all) based on probability densities and to add class models without affecting the boundaries of existing models. Experiments show that for several classification problems using class-specific models significantly improves the performance.