Fuzzy Logic-Based Scenario Recognition from Video Sequences

Abstract

In recent years, video surveillance and monitoring have gained importance because of security and safety concerns. Banks, borders, airports, stores, and parking areas are the important application areas. There are two main parts in scenario recognition: Low level processing, including moving object detection and object tracking, and feature extraction. We have developed new features through this work which are RUD (relative upper density), RMD (relative middle density) and RLD (relative lower density), and we have used other features such as aspect ratio, width, height, and color of the object. High level processing, including event start-end point detection, activity detection for each frame and scenario recognition for sequence of images. This part is the focus of our research, and different pattern recognition and classification methods are implemented and experimental results are analyzed. We looked into several methods of classification which are decision tree, frequency domain classification, neural network-based classification, Bayes classifier, and pattern recognition methods, which are control charts, and hidden Markov models. The control chart approach, which is a decision methodology, gives more promising results than other methodologies. Overlapping between events is one of the problems, hence we applied fuzzy logic technique to solve this problem. After using this method the total accuracy increased from 95.6 to 97.2.

A Fast and Robust Method for Image Segmentation Using Fuzzy Solutions of Partial Differential Equations

“…The unknown solution of the Poisson equation for image segmentation with the BCs in (5) can be regarded as an unknown system, and it can be discretized and solved in a number of ways. And the fuzzy logic system is very suitable for treating this problem with its excellent approximation ability [18].…”

Implementation of NOGIE and NOWGIE for Human Skin Detection

“…First is to acquire the optimum results the RGB space is converted into the HSV space, HSV color space is independent of the three-color components of the RGB color space. Then calculate the differences in the Histograms of successive video frames on each color component respectively the calculated Histogram difference is the base for the feature detection, this fuzzy logic based human skin detection technique is proved to be very satisfactory, but the Interpolation methodology have proven to be far better [6], [7]. Applying the skin segmentation along with the fuzzy logic is another technique where the whole frame comprising skin pixels is divided into two parts one is the fuzzy part [6] and the other is skin segmentation, training the system requires images with face and non-face here the image pixels are read in row segments to form the column segments the images consisting of face the system takes then as 1 while the comprising of no face the system takes it as a T-S model based fuzzy training is implemented and the fuzzy function comprising weights are learnt via algorithm training the system requires multiple images of different pixel values, distances and sizes, 69 face and 56 faceless images were provided for system training for accurate results in detecting faces at instant speed [8].…”

“…Then calculate the differences in the Histograms of successive video frames on each color component respectively the calculated Histogram difference is the base for the feature detection, this fuzzy logic based human skin detection technique is proved to be very satisfactory, but the Interpolation methodology have proven to be far better [6], [7]. Applying the skin segmentation along with the fuzzy logic is another technique where the whole frame comprising skin pixels is divided into two parts one is the fuzzy part [6] and the other is skin segmentation, training the system requires images with face and non-face here the image pixels are read in row segments to form the column segments the images consisting of face the system takes then as 1 while the comprising of no face the system takes it as a T-S model based fuzzy training is implemented and the fuzzy function comprising weights are learnt via algorithm training the system requires multiple images of different pixel values, distances and sizes, 69 face and 56 faceless images were provided for system training for accurate results in detecting faces at instant speed [8]. A simple method is applied for the acquisition of skin pixels from RGB images consisting of facial constraints, such RGB image is taken as an input then techniques are applied for the detection of nose, the color pixels of nose's skin tone is extracted the nose is considered as a main region for the identification of the same skin pixels from the facial regions after that skin segmentation is performed and histogram model is constructed by applying fusion strategy via Gaussian Model, the results are obtained and compared both Gaussian model and fusion strategy gives good results but the Fusion strategy provides the best output [9].…”

Utilizing Multimedia Ontologies in Video Scene Interpretation via Information Fusion and Automated Reasoning

“…Fuzzy description logics can be used to express the certainty of the depiction of concepts [24], events, and video scenes [25]. This can be achieved by enabling normalized certainty degree values assigned to objects of fuzzy concepts.…”