Human Behaviour Analysis using Fuzzy Neural Networks

Submission Deadline: 17 June 2023 Submit to Special Issue

Guest Editors

Dr. M.M. Kamruzzaman, Jouf University, Kingdom of Saudi Arabia
Dr. Shuai Liu, Hunan Normal University, China
Dr. Muhammad Aslam Jarwar, Sheffield Hallam University, United Kingdom

Summary

Understanding human characteristics and behavioural cues are the keys to a lot of futuristic technologies that aim at improving security and surveillance. It becomes an increasingly significant technology since identifying dangerous situations through visual information can be very effective in averting accidents and crimes. The need for intelligent systems and fuzzy neural networks becomes prominent due to the inherent inadequacy in analysing large sets of data by human resources. Henceforth a lot of algorithms and innovative frameworks are being tested and used across various applications to aid in human behaviour analysis quickly and accurately.

 

The various areas in which human behaviour analysis using fuzzy neural networks can be applied includes a prediction of the activities and actions of the user, energy usage modelling, pathology stratification, crime detection etc. All these applications follow a similar theory: the future evolution of user behaviours or detection of behaviour/conduct deviation can be assessed against the data using intelligent algorithms. A few of the techniques involved are as follows: gaussian mixture regression and gaussian mixture modelling for recognising motion primitives, intelligent big data analytics for analysing human behaviour using large amounts of social media data, decision trees, Support Vector Networks (SVNs), Naive Bayes, Recurrent Neural Networks (RNNs), Convolutional Neural Networks (CNNs) etc. In recent days, fuzzy neural networks have been showing excellent results in analysing macro behaviours of humans while a time delay neural network is used to analyse micro human behaviours. This type of architecture follows a hierarchy among the techniques used. Such innovation will highly improve the scalability, effectiveness, and robustness of the currently available human behavioural analysis systems. These mechanisms will significantly avoid the imprecision involved in tracking and work based on semantic analysis and context-aware algorithms. A hierarchy-based human behaviour analysis algorithm using fuzzy neural networks will consist of a combination of the following technologies: computer vision, knowledge representation and reasoning, and probabilistic reasoning. Another interesting technique follows a novel fuzzy machine vision-based algorithm. Slices of silhouette and speed of movements are extracted using an efficient model-based feature set from video inputs. By leveraging the fuzzy c-means clustering algorithm, the recognised behaviour is taken as the highest degree of output from the candidate’s best behaviour dataset.


Keywords

Recognition of suspicious human activities using fuzzy rule inference algorithms.
Novel biologically inspired human behaviour analysis algorithm using social media data.
Developing a fuzzy logic-based model to answer questions about moving targets in a video surveillance system.
Innovative action classification to analyse human behaviour using discriminative subsequence mining.
Human action and behaviour analysis using pose primitives for crime detection.
Methods to test the validity of clustering mechanisms in a fuzzy c-means-based human behaviour analysis model.
Innovative fuzzy logic methods for detection of irregular activities in crowded videos.
Information maximization for human behaviour analysis.
Design and development of intelligent fuzzy logic-based systems for uncertain occurrences.
Development of a fuzzy logic-based human behaviour and action recognition model for elderly care.

Share Link

WeChat scan