COMPUTATIONAL MODEL FOR CREATING NEURAL NETWORK DATASET OF EXTRACTED FEATURES FROM IMAGES CAPTURED BY MULTIMEDIA SECURITY DEVICES

Abstract

Whenever multimedia security devices, such as the Closed-Circuit Television (CCTV), capture images, the decisive analysis is usually left for the human expert to determine the content therein and suggest the necessary action to be taken. However, the application of Artificial Intelligence (AI) helps in complementing human efforts in carrying out such analysis. This is achievable if the required input dataset is created by extracting features from the identifiable objects in the images. The extraction of such features is based on regional properties of objects within an image – using technique such as the Gray-Level Co-occurrence Matrix (GLCM). This dataset is consequently used in AI platforms that are based on Artificial Neural Network (ANN) and Neuro-Fuzzy systems (specifically in Adaptive Neuro-Fuzzy Inference System (ANFIS)). This paper is presenting a computational model for creating training and testing datasets. For the simulation of the model, Matlab was used, however, the computational model is realizable via other programming and numerical computing environments.