Khaled Mohamed Saad
Mai Magdy El-Ghandour
Ahmed Mohamed Tamer
Reem Ahmed Ibrahim
Assoc.Prof. Eslam Amer
Eng. Mahmoud Elsahhar

Publishing Date

5 January 2022


Movies and series with different perspectives have become very important and play an important role in youngsters’ lives. but be filled with inappropriate content such as violence and sexual scenes may harm them, Researches has shown that crime rates have increased by people within the age of 12 to 17 as 92 per 1,000 youngsters commit a crime. The research conducted as part of the government’s troubled attempt to introduce age-verification services for online pornography found that children were often disturbed by being accidentally exposed to the material at a pre-teenage. There are many faced challenges in developing filtering systems for negative content. Most researches on negative content filtering have been based on video analysis only, Our main target is to merge between video analysis on scenes parallel to text analysis on subtitles to predict the violent and inappropriate scenes from movies and cut them off.

1.1 Purpose of this document

This document is purposed to define the details of the Smart Violence Pruning System from the Movies document. Additionally, it will explain the route of implementation of the software. This software implementation includes the used methods and algorithms in the system. Smart Violence Pruning System will assist people who would like to remove inappropriate scenes from movies to create acceptable content. Therefore, the system will predict the violence of the next scenes. After that, it will filter these scenes according to the prediction to make them appropriate. The user will upload a video as input. Continuously, the system will filter the taken video. According to the subtitles classification and the frames of its image to compares with fixed datasets used to detect unsuitable scenes.

1.2 Scope of this document

The Pruney system allows people to watch all movie types with satisfaction. To make them sure enough that the movie doesn’t contain any unsuitable scenes makes them annoyed. Important to realize that users will deal with a system able to remove unwanted content from movies. To clarify that the system can know the current scene and predict the upcoming scenes of the movie. Thus, it can remove or blur the violent contents depending on user choice. Users do not need to leave their movie till finishing cutting or blurring. In detail, the filtration and prediction processes work while watching the movie. These prove the system’s high efficiency in detecting inappropriate scenes.

1.3 System Overview

A desktop application that aims to produce clean videos or movies without inappropriate scenes. The application cuts these scenes using prediction, the system predicts the next scene and decides if it’s inappropriate or not using our Markov models that are used to predict the sequence of emotions of the inappropriate scenes based on the pre-violence scene emotions and based on the comparison with the markov model of the violence dataset and markov model with the non violence dataset, the system decide if this scene is violent or not. The prediction is based on a combination of text and video emotion analysis, both of them are used in the prediction. The user will upload his preferred video or movie and he will choose which type of inappropriate scenes he wants to cut, also the user can choose how to filter his input either by cutting or blurring it. Finally, after filtering the user’s input it will be saved with the new clean version of it. The user could watch the new version anytime later without repeating the whole process again.

1.4 System Scope

• Extracting emotions from video frames.

• Extracting emotions from subtitles and sound.

• Video and Text analysis.

• Predicting the violent and inappropriate scenes.

• Cutting or Blurring the violent and inappropriate scenes.

• Categorizing the filtered movies.

The expected outcome of the project is a usable system that can filter movies from any inappropriatescenes according to the user’s choice and save it to his account and add to a certain category.