WOA-FNN: An innovative hybrid optimization technique for effective detection of shot boundaries
Tushar Banik, Saptarshi Chakraborty, Dalton Meitei Thounaojam, Debjani Bhowmik · 2023
The presented research proposes a unique approach for Shot Boundary Detection (SBD) of videos using the Whale Optimisation Algorithm (WOA) to identify both abrupt and gradual transitions. The Whale Optimization Algorithm optimizes the weights and bias of the Feed-Forward Neural Network (FNN) to classify transition frames and normal frames from a set of available frames, with the aid of an experimental threshold. WOA has a unique characteristics of fast convergence and searching for optimal solutions with high precision. The proposed system has adopted a dual stage system comprising of possible and confirmation stage to increase the effectiveness of the system interms of F1 Score. WOAFNN gives a better results interms of precision and F1 score when compared with the state-of-art techniques.