In this paper we present a new framework for novelty detection. The framework evaluates neural networks as adaptive classifiers that are capable of novelty detection and retraining on the basis of newly discovered information. We apply our newly developed model to the application area of object recognition in video. The application is however not limited to scene analysis and the basic methodology can be easily extended to other areas. This paper details the tools and methods needed for novelty detection such that data from unknown classes can be reliably rejected without any a priori knowledge of its characteristics. The rejected data is post-processed to determine which samples can be manually labelled of a new type and used for retraining. In this paper we compare the proposed framework with a naïve solution and discuss the results of retraining neural network to recognise further unseen data containing the newly added objects.
|
89
|
Pattern Classification, 2nd Edition
– Duda, Hart, et al.
- 2001
|
|
60
|
Novelty Detection and Neural Network Validation
– Bishop
- 1994
|
|
54
|
Support vector domain description
– Tax, Duin
- 1999
|
|
49
|
Support vector method for novelty detection
– Schölkopf, Williamson, et al.
- 1999
|
|
43
|
A novelty detection approach to classification
– Japkowicz, Myers, et al.
- 1995
|
|
39
|
Novelty detection for the identification of masses in mammograms
– Tarassenko, Hayton, et al.
- 1995
|
|
38
|
A Linear Programming Approach to Novelty Detection
– Campbell, Bennett
- 2001
|
|
34
|
FCM: the fuzzy C-mean clustering algorithm
– Bezdek, Ehrlich, et al.
- 1984
|
|
32
|
Statistical independence and novelty detection with information-preserving nonlinear maps
– Parra, Deco, et al.
- 1996
|
|
25
|
Objective functions for training new hidden units in constructive neural networks
– Kwok, Yeung
- 1997
|
|
17
|
Novelty, confidence and errors in connectionist systems
– Roberts, Penny, et al.
- 1996
|
|
13
|
Tarassenko L. A system for the analysis of jet engine vibration data
– Nairac, Townsend, et al.
- 1999
|
|
13
|
Structural fault detection using a novelty measure. Journal of Sound and Vibration, 201(1):85 – 101
– Worden
- 1997
|
|
11
|
Novelty Detection Using Extreme Value Statistics
– Roberts
- 1999
|
|
10
|
A real-time novelty detector for a mobile robot
– Marsland, Nehmzow, et al.
- 2000
|
|
10
|
Detecting novel features of an environment using habiutation
– Marsland, Nehmzow, et al.
- 2000
|
|
10
|
Tarassenko L. Choosing an Appropriate Model for Novelty Detection
– Nairac, Corbett-Clark, et al.
- 1997
|
|
9
|
Classification and Novelty Detection Using Linear Models and a Class Dependent-Elliptical
– Brotherton, Johnson, et al.
- 1999
|
|
9
|
A robot implementation of a biologically inspired method for novelty detection
– Crook, Hayes
- 2001
|
|
8
|
A novelty detection approach to diagnose damage in a cracked beam
– Surace, Worden, et al.
- 1997
|
|
7
|
Generalised radial basis function networks for classification and novelty detection: self-organisation of optimal Bayesian decision
– Albrecht, Busch, et al.
- 2000
|
|
7
|
Certain principles of biomorphic robots”, Autonomous Robots (in press
– Lewis
- 2001
|
|
7
|
Novelty detection in large environments
– Marsland, Nehmzow, et al.
- 2001
|
|
6
|
A Model of Habituation Applied to Mobile Robots
– Marsland, Nehmzow, et al.
- 1999
|
|
6
|
Novelty detection for robot neotaxis
– Marsland, Nehmzow, et al.
- 2000
|
|
6
|
Novelty detection using products of simple experts- a potential architecture for embedded systems
– Murray
- 2001
|
|
6
|
Evident: a functional magnetic resonance image analysis system
– Pizzi, Vivanco, et al.
- 2001
|
|
6
|
Novelty detection in jet engines
– Tarassenko, Nairac, et al.
- 1999
|
|
5
|
Applications of probability density estimation to the detection of abnormal conditions in engineering
– Desforges, Jacob, et al.
- 1998
|
|
4
|
An immunogenetic approach to intrusion detection
– Dasgupta, Gonzalez
- 2001
|
|
4
|
Improving the performance of the radial basis function classifiers in condition monitoring and fault diagnosis applications where “unknown” faults may occur”, Pattern Recognition Letters (in press
– Li, Pont, et al.
- 2002
|
|
4
|
The importance of being emergent
– Saunders, Gero
- 2000
|
|
3
|
Detection of shorted-turns in the field of turbinegenerator rotors using novelty detectors- development and field tests
– Streifel, Maks, et al.
- 1996
|
|
3
|
A novelty detection method to diagnose damage in structures: an application to an offshore platform
– Surace, Worden
- 1998
|
|
3
|
Recognizing novelty in classification tasks
– Vasconcelos, Fairhurst, et al.
- 1994
|
|
3
|
Clustering of the self-organising Map
– Vesanto, Alhoniemi
- 2000
|
|
2
|
Minerva scene analysis benchmark
– Singh, Singh
- 2001
|
|
2
|
Detection of new image objects in video sequences using neural networks
– Singh, Markou, et al.
- 2000
|
|
2
|
Novelty detection using auto-associative neural network
– Sohn, Worden, et al.
- 2001
|
|
2
|
A bootstrap-like rejection mechanism for multilayer perceptron networks
– Vasconcelos
- 1995
|
|
1
|
Neural network
– Harris
|
|
1
|
NOMAD: a documentary database interrogation system using multiple neural topographies and novelty detection
– Lamirel, Crehange, et al.
- 1994
|
|
1
|
Application of two damage detection techniques to an offshore platform
– Ruotolo, Surace, et al.
- 1999
|
|
1
|
Recent advances in distributed collaborative surveillance
– Saptharishi, Bhat, et al.
- 2000
|
|
1
|
Some aspects of novelty detection methods
– Surace, Worden
- 1997
|
|
1
|
Improved classification for a data fusing Kohonen self organising map using a dynamic thresholding techniques
– Taylor, Tait, et al.
- 1999
|
|
1
|
Novelty detection using self-organising maps”, Progress in Connectionist Based Information Systems
– Ypma, Duin
- 1998
|
