Auto-Censor: Control What You See

Auto-Censor: Control What You See Amina Khatun Mim∗ , Israt Zaman∗ , Kazi Sinthia Kabir∗ , and Mayisha Alam∗ ∗ Department of Computer Science and Engineering Bangladesh University of Engineering and Technology, Dhaka, Bangladesh Email: {amina.k.mim12, isratzaman99, sinthia.096,mayisha007}@gmail.com Author names are given in alphabetical order. 90

Violent actions in film

80 70 60

18.5

50

44.1

40

Boys

lethal force

37.4

Girls

30

moderate force minimal force

20 10 0 Have seen child Have seen Have seen pornogrphy bestiality online sexual bondage online

Fig. 1.

Percentage of violent actions in movies

Abstract—Obscene videos,moviesand any other entertainment forms are causing a great destruction in our society. At this present era of internet , it’s not possible to check using internet or watching HD videos but to refine the videos or check internet contents what the auto censor will do. The system will work according to “LiveLight” method. It would continuously compare the scenes of videos with a given dictionary contents and thus detect obscenes and then cut off these scenes.

I.

I NTRODUCTION

Obscenity is considered as one of the biggest causes of degradation of morality and social values in the society. At present, movies and videos on the internet have become a great source of entertainment. Therefore, obscene videos are being seen frequently intentionally or unintentionally. Previous research shows us that if we are able to control obscene videos then we might be able to save our generation to some extent. We are motivated by the circumstances to think of a system named ‘auto censor’ which will help to check obscenity. II.

BACKGROUND AND R ELATED W ORK

One of the main problem in the present world are exposed to obscene videos. This problem is affecting our social life very badly. A research was done over 100 films released in 1994 in Hollywood. A total of 2184 violent actions against people were identified in those films. Even one of the family films (Lion King, Walt Disney Productions) had nearly as many97 separate violent actions. The data from research show that among 2184 violent actions 963 (44.1%) violent actions used lethal force, 817 (37.4%) used moderate force, and 404 (18.5%) used only minimal force. Lethal force was most frequent in the action genre, in which 56.8% of all violent actions were rated lethal. Explicitness was rated either none or minimal in 1932 scenes (88.4%) but moderate in 222 (10.2%) and maximum in 30 (1.4%). Included in this last category were scenes from seven of the 14 action films. In short the result of the research is The median number of violent actions per film was 16, with a range from 1 to 110. Intentional violence outnumbered unintentional violence by a factor of 10. Almost 90% of violent actions showed no consequences to the recipient’s body, although more than 80% of the violent actions were executed with lethal or moderate force. Fewer than 1% of violent actions were accompanied by injuries that were then medically attended [1].

Fig. 2.

Have seen group sex online

Have seen same-sex intercourse online

Average number of boys and girls who watched obscene videos

Another statistics shows that, on average: 6 out of 10 girls were exposed to pornography before the age of 18.15% of boys and 9Exposure to obscene videos and “steady diet of violent content over time” develop callousness toward women and trivialize rape and violent act as a criminal offense; to some it was no longer acrime at all [2]. This develop distorted perceptions about sexuality and an appetite for more deviant, bizarre, or violent types ofpornography (escalation); normal sex no longer seemed to “do the job”- [3]. Continuous consumption of obscene videos devalue the importance of monogamy and view non-monogamous relationships as normal and natural behavior. It affects children the most. While watching televisionand movies, children identify themselves with certain characters. Male children often identify with male super heroes who use violence. They quickly learn that violence is an acceptable solution to resolving even complex problems, particularly if the aggressor is the hero. According to the American Academy of Child Adolescent Psychiatry: The average American child will view more than 200,000 acts of violence before the age of 18. A study in the journal “Pediatrics” found that R-rated movies were being watched by more than 12 percent of American children aged between 10 and 14.A study by Albert Bandura in 1965 concluded that the more onscreen violence a child is exposed to, the greater the chance that they would engage in aggressive behavior [4]. “Not every child who watches a lot of violence or plays a lot of violent games will grow up to be violent. Other forces must converge, as they did recently in Colorado. However, just as every cigarette increases the chance that someday you will get lung cancer, every exposure to violence increases the chances that some day a child will behave more violently than they otherwise would.” - Ibid attributed to L. Rowell Huesmann of the University of Michigan at Ann Arbor. Over 1000 studies - including a Surgeon General’s special report in 1972 and a National Institute of Mental Health report 10 years later - attest to a causal connection between media violence and aggressive behavior in some children. Studies show that the more “real-life” the violence portrayed, the greater the likelihood that it will be “learned.”- American Academy of Pediatrics Policy Statement, Volume 95, Number 6 - June 1995. At present there is some system for editing or filtering obscene videos . The mostly used systems are video editing and filtering. Video

A. Comparing with the dictionary Auto Censor is a system that would be designed based on the working principle of Livelight. The only difference is that the algorithm of Auto Censor will not compile a dictionary; instead it will compare the segments with the contents of a system dictionary. There would be three individual dictionaries comprising our whole system dictionary in the algorithm. The dictionaries would already contain some specific contents to be regarded as obscene. While evaluating the whole video each segment would be compared with the dictionary contents. Fig. 3.

Overview of the auto-sensor system : dictonary of sample scene

B. Identify and remove obscene Evaluating the whole video the system would identify the obscene working as following:

Fig. 4.

Overview of auto-censor system: elimination of obsceinity

editing is the process of manipulating and rearranging video shots to create a new work. If someone wanted to cover a section of the video, the overlay function of these programs could be used. A user could cover with a picture, or even a video. A user could place a black bar or blur over the video track. Filtering technology consists of software that screens out some content while allowing other material to flow through to its intended destination. Filtering technology comes in several forms: client side server, content limits or content filters from ISPs, server side filters, search engine filters, etc [5]. However, these systems have so many pitfalls.The video editing systems presented here are not automatic and efficient. The user has to do some manual work to edit the video and cut off obsence from the video.The filter technology is not perfect - some desirable material may be accidentally blocked and some objectionable material may slip through the cracks . While effective in screening out accidental “hits” or results that include inappropriate content, search engine filters do not restrict access to content if a surfer directly types in a URL [5]. III.

U NIQUENESS OF THE S YSTEM

“Auto Censor” provides a system to free Videos from obscenity. This project provides a system which needs no manual operation. Being completely an automatic system, it would bring more efficiency, saves time, ensure accuracy as well as overcome the problems related to filtering. IV.

OVERVIEW O F T HE S YSTEM

This system will use the “Livelight” method [6] (Livelights main function is to create the summary of a video by eliminating the repeated scenes) to remove obscene parts from the videos. It might also be used to prevent sharing such videos over the Internet. Comparing the scenes with the contents of a system dictionary it would detect obscene.

•

There would be a set of dialogues in the first dictionary containing filthy words. A scene where any character uses similar dialogue would be removed.

•

Then the actions or movements which are uncensored would be detected by including those kinds of scenes in the second dictionary and the system would finally exclude the detected scene from the video.

•

The third one would contain some images to be the standard of uncensored dress-up. Any scene with a character wearing less than the standard would be marked as obscene. Auto Censor would evaluate the whole movie and eliminate the uncensored segments.

C. Censored video The final output of the system would be a censored video which contains no obscenity and might be shorter in length in case of finding obscene and eliminating those. V.

F UTURE W ORK

The movies or videos created using the existing video recorders or cameras can be made censored by using “Auto Censor” by setting up the application in the cameras. Also the newer ones would contain this software. As the dimension of obscenity is changing day by day the system would need update in it’s dictionary including new contents. So an updated version of the application would be introduced after a certain period of time regularly. However, if this software could be included into the cameras as a built in software and the system could be converted to work while recording, there would be no obscenity present in any video. Because it will then become impossible to record any uncensored scene using the camera. VI.

CONCLUSION

Focusing the circumstances of present time we can illustrate that the ‘auto censor’ system we have implemented above can help the society to a great extent. Moreover, it can also be developed and implemented with some device in future to make it more useful. R EFERENCES [1] [2] [3] [4]

“Injury prevention,” http://injuryprevention.bmj.com/content/6/2/120.full. “Ibid - attributed to kathryn c. montgomery, president of the center for media education,” “Study conducted by rand and published in the september 2004 issue of pediatrics,” “Effect of violent movies on kids,” http://www.ehow.com/info 8357089 effect-violencemovies-kids.html. [5] “Online safety,” http://www.justice.gov/criminal/ceos/additionalresources/onlinesafety.html. [6] “Automatic cutting of boring parts from long videos,” http://www.sciencedaily.com/releases/2014/06/140625132441.htm.