JOURNAL OF INTELLIGENT SYSTEMS WITH APPLICATIONS

Year: 2018, Volume: 1, Number: 2
Published : Jan 29, 2026

Anomaly Detection with Structural Similarity and Optical Flow Histogram

Kadriye Öz (1), İsmail Rakıp Karaş (2)

(1) Mekatronik Mühendisliği Bölümü, Karabük Üniversitesi, Karabük
(2) Bilgisayar Mühendisliği Bölümü, Karabük Üniversitesi, Karabük
Fulltext View | Download
Abstract

In this paper, we present an anomaly detection and localization system for surveillance systems. A new feature descriptor is proposed. The spatio-temporal identifiers are obtained by using optical flow histogram and the structural similarity index from the videos that contain normal conditions. An artificial neural network, Selforganizing maps are used in modeling. The proposed system has been tested on the UCSD dataset.

Managed by Open Journal System