A Hybrid Video Shot Boundary Detection

This system formulates a novel Shot Boundary Detection (SBD) approach by optimizing the biases and weights of the Feed forward Neural Network (FNN) using the Advanced Grasshopper Optimization Algorithm (AGOA). The way grass hopper swarms navigate in the na-ture served as the model for AGOA. A mathematical model is used to simulate how members of the swarm interact, taking into account factors like comfort zone, attraction, and repulsion. It detects the boundary of a video shot and improves system performance. To retrieve the maxi-mum number of transition frames, an outlier and a continuity matrix Φ are utilized. To identify shot boundaries, a set of thresholds δ1 and δ2 is chosen in order to distinguish between sudden and smooth transition frames. The TRECVid 2001 and TRECVid 2007 datasets are used for the experiment, which shows that the Advanced Grasshopper Optimiza-tion Algorithm Feed forward Neural Network (AGOA-FNN) is capable of accurately detecting both rapid and gradual shot transitions. The out-come demonstrates that, in comparison to other recent tests conducted in recent years, the suggested model can produce results that are highly competitive in terms of accuracy

Artificial neural network · Outlier · Particle swarm optimization · Anomaly Detection Techniques and Applications · Human Pose and Action Recognition · Video Analysis and Summarization