In this research, I would like to introduce a novel abnormality detection method using image prediction. The technique involves two main steps: extracting the principle images of the animation using Principle Component Analysis (PCA), then followed by image prediction using Signal Spectrum Analysis (SSA). The algorithmic combination of PCA followed by SSA successfully predicted future video animation images with little error (<1%) when compared to the original. This technology can potentially aid prediction and detection efforts of security systems.
doi: 10.12783/SHM2015/142