One-Class Classification with Gaussian Processes
نویسندگان
چکیده
Detecting instances of unknown categories is an important task for a multitude of problems such as object recognition, event detection, and defect localization. This paper investigates the use of Gaussian process (GP) priors for this area of research. Focusing on the task of one-class classification for visual object recognition, we analyze different measures derived from GP regression and approximate GP classification. Experiments are performed using a large set of categories and different image kernel functions. Our findings show that the well-known Support Vector Data Description is significantly outperformed by at least two GP measures which indicates high potential of Gaussian processes for one-class classification.
منابع مشابه
Negative Selection Based Data Classification with Flexible Boundaries
One of the most important artificial immune algorithms is negative selection algorithm, which is an anomaly detection and pattern recognition technique; however, recent research has shown the successful application of this algorithm in data classification. Most of the negative selection methods consider deterministic boundaries to distinguish between self and non-self-spaces. In this paper, two...
متن کاملMulti-class Classification with Dependent Gaussian Processes
We present a novel multi-output Gaussian process model for multi-class classification. We build on the formulation of Gaussian processes via convolution of white Gaussian noise processes with a parameterized kernel and present a new class of multi-output covariance functions. The latter allow for greater flexibility in modelling relationships between outputs while being parsimonious with regard...
متن کاملDivergence-Based One-Class Classification Using Gaussian Processes
We present an information theoretic framework for one-class classification, which allows for deriving several new novelty scores. With these scores, we are able to rank samples according to their novelty and to detect outliers not belonging to a learnt data distribution. The key idea of our approach is to measure the impact of a test sample on the previously learnt model. This is carried out in...
متن کاملA Family of Non-Gaussian Martingales with Gaussian Marginals
We construct a family of martingales with Gaussian marginal distributions. We give a weak construction as Markov, inhomogeneous in time processes, and compute their infinitesimal generators. We give the predictable quadratic variation and show that the paths are not continuous. The construction uses distributions Gσ having a logconvolution semigroup property. Further, we categorize these proces...
متن کاملOne-Class Classification for Anomaly Detection in Wire Ropes with Gaussian Processes in a Few Lines of Code
Anomaly Detection in Wire Ropes is an important problem. Detecting suspicious anomalies in the rope surface is challenging because of the variety of its visual appearance caused by reflections or mud on the rope surface. This hinders the discrimination between uncritical variations and small defects within the rope surface enormously. The fact that nearly no defective samples are available to t...
متن کامل