Weakly Supervised Learning of a Classifier for Unusual Event Detection

Abstract

In this paper, we present an automatic classification framework combining appearance based features and hidden Markov models (HMM) to detect unusual events in image sequences. One characteristic of the classification task is that anomalies are rare. This reflects the situation in the quality control of industrial processes, where error events are scarce by nature. As an additional restriction, class labels are only available for the complete image sequence, since frame-wise manual scanning of the recorded sequences for anomalies is too expensive and should, therefore, be avoided. The proposed framework reduces the feature space dimension of the image sequences by employing subspace methods and encodes characteristic temporal dynamics using continuous hidden Markov models (CHMMs). The applied learning procedure is as follows. 1) A generative model for the regular sequences is trained (one-class learning). 2) The regular sequence model (RSM) is used to locate potentially unusual segments within error sequences by means of a change detection algorithm (outlier detection). 3) Unusual segments are used to expand the RSM to an error sequence model (ESM). The complexity of the ESM is controlled by means of the Bayesian Information Criterion (BIC). The likelihood ratio of the data given the ESM and the RSM is used for the classification decision. This ratio is close to one for sequences without error events and increases for sequences containing error events. Experimental results are presented for image sequences recorded from industrial laser welding processes. We demonstrate that the learning procedure can significantly reduce the user interaction and that sequences with error events can be found with a small false positive rate. It has also been shown that a modeling of the temporal dynamics is necessary to reach these low error rates.

Artificial Cognition in Production Systems

“…Any variation in process parameters often necessitates a recalibration, which might even require the assembly line to be halted. Research has offered promising approaches for controlling laser welding processes [138] and for teaching machines how to handle them [139]- [141]. In a Cognitive Factory, many different sensors are combined and real-time machine-learning techniques to improve process quality and control are developed.…”

Dynamic Bayesian Network-Based Anomaly Detection for In-Process Visual Inspection of Laser Surface Heat Treatment

“…As proposed by other visual inspection methods, the analysis of these emissions will help to detect defects during laser processes [3], [15].…”

“…This is an obstacle to training automated systems with statistical learning models [3]. One-class classication is the anomaly detection technique used in machine learning for binary classication when information on only one of the classes is available.…”

Behavior tracking model in dynamic situation using the risk ratio EM

“…This section provides a recent review of the works closely related to object detection, human behavior recognition, with machine learning and probabilistic model. The purpose of human behavior detection is to recognize, or learn interesting events, which is defined as "suspicious event" [3], "irregular behavior" [4], "uncommon trajectory" [5], unusual activity/behavior [6][7][8][9][10][11][12][13], "abnormal behavior" [14][15][16] to predict dangerous situation for pedestrians. For the improvement of public safety level and prevention of the crime, intelligent real-time human behavior recognition is more important in daily life.…”