Paper Title
DEEP LEARNING-BASED FILTERING AND CLASSIFICATION OF IN APPROPRIATE CONTENT IN YOUTUBE VIDEOS

Abstract
The rapid growth of video content on digital platforms has raised significant concerns about inappropriate visuals, such as violence and explicit content. This work introduces a new advanced deep learning architecture framework for detecting and classifying inappropriate video content. The system combines a pre-trained EfficientNet-B7 Convolutional Neural Network (CNN) for spatial feature extraction with a Bidirectional Long Short- Term Memory (Bi- LSTM) network for temporal analysis. The combined technique well captures both the spatial and temporal relations, supporting correct classification. Experimental results show that the suggested model attains 94.6% accuracy, outperforming existing methods. This framework can significantly reduce manual content moderation efforts, improving safety andefficiency on video platforms [1][2]. Keywords - DeepLearning, EfficientNet-B7, Bi-LSTM, Inappropriate Content Detection, Video Classification, Content Moderation.