Snitch Plus is designed to detect inappropriate material of various forms on a computer's hard drives. It uses a number of different porn detection techniques to isolate adult content from other files, including analysing image and video files for nudity, detecting suspicious key phrases in filenames, text-based files, Internet history, file Metadata and other sources. It is also capable of finding suspicious items hidden in various ways on the hard disk. We will discuss some of these key techniques in this document.
Snitch Plus uses proprietary SkinSkan™ technology to scan image and video files, and automatically detect the presence of nudity or adult content. The unique SkinSkan™ technology in Snitch Plus analyses visual content for key elements that indicate the presence of porn, including skin colour, tone levels and distribution, body shapes, faces, and a range of other important factors. These factors are then used to generate a probability profile that is compared to precompiled statistical samples taken from a huge database of known pornographic sources. This process produces a suspicion rating which is used as a baseline for further testing of that file. Subsequent tests can modify this baseline value depending on additional indicators. A few of these will be discussed next.
Snitch Plus incorporates a robust keyphrase analysis module that is used to test any available text components of the file being analysed. For example, filenames are a good indicator (when used in conjunction with other tests) of possible pornographic content in files. Some target files contain metadata that can also be analysed for keyphrase content. These supplementary checks apply either a positive or negative weighting factor to the baseline suspicion rating, effectively reducing or increasing the total suspicion value that file, depending on the context of the key phrases, the individual weightings of the key phrases detected, and the presence of 'negative keywords' and other factors. The baseline suspicion rating for this file is then recorded and used to rank it against all the other detected items when presenting the results of the search.
Snitch Plus uses a variety of techniques to detect inappropriate files that have been hidden intentionally. For example, sometimes users will change the file extension of a file to change the file association, with the expectation that casual observers will not realise it is a pornographic file. Suppose a user had downloaded a pornographic video called "myfile.mpg", and then tried to disguise its true nature by changing its extension to "myfile.txt". Double-clicking that file would then invoke the default text editor, and fail (since the file is not a text file); and the file would not appear with a video icon in Windows Explorer either. This 'disguising' process makes it harder to find the file. But not for Snitch Plus.
It is able to analyse a file's binary format to determine what its actual file type is, so even if it has been renamed as in the example, it will still be detected as a video file. In results it will appear in the renamed file category section, where it can be tested further if necessary, by viewing it as a video file.
Snitch Plus is also able to look inside file containers such as compressed archives (ZIP, RAR, etc.) to find images and videos stored there that may be suspicious.
Inappropriate Internet history detection and cleaning is another area where Snitch Plus excels. It can rapidly check browser cache files and Internet history logs to detect visitation of inappropriate Internet URLs such as adult entertainment and porn sites. Snitch Plus analyses not only the root domain name, but also all the visited page names and associated embedded files, to determine if they are porn related. This analysis is useful in establishing a pattern of abuse of Internet resources.
How terms like 'pron' and porn impact porn detection effectiveness
The legal implications of when Snitch detects illegal porn during a porn scan
Configuring Snitch Plus pornography cleaner to remove browser history
Improving system performance by erasing detected porn files
Porn Detection vs Pokemon style porn files - a case study