Abstract:

YouTube's video content has grown exponentially, drawing billions of viewers, the vast majority of whom are young. Additionally, malicious uploaders use this site to disseminate disturbing visual content, such as improper content for minors via animated cartoon videos. Therefore, it is strongly advised that social media networks have an automated real-time video content screening method. This paper suggests a revolutionary deep learning-based architecture for identifying and categorising films that include unsuitable information. A pre-trained convolutional neural network (CNN) model called EfficientNet-B7 from ImageNet is used in the suggested framework to extract video descriptors. These are subsequently fed into a bidirectional long short-term memory (BiLSTM) network, which learns efficient video representations and performs multiclass video classification

Description:

.