Link
What is this?
SmutDetect is a skin-tone image mining software which scans directories for images containing a specified percentage of skin-tones. It ranks (and filters) these images in various reports. Hash-export can be used to process the results in TSK/autopsy.
What I want
I would be more than happy if you would take the time to test this software and let me know about bugs via the bug tracker on SF or email. Also I would be more than pleased if you have further suggestions for the improvement of this application.
Background Info
During my current studies of Forensic Computing I started writing the application SmutDetect which is supposed to speed up the findings of skin tone images on a suspect storage device / recovered digital evidence. This skin tone image mining software is running quite stable at the moment but is still in Beta stage. The algorithms behind it are NOT new as they have been taken from the documented concept of FileHound. (See Code comments for more details.) FileHound though is not easily obtained as it is only given to "select academic and research organizations" and law enforcement.
SmutDetect provides a free (in the understanding of the GPL) implementation of established algorithms trying to aid the digital forensic investigator. Currently the implementation is limited to RGB and YCbCr detection algorithms. A GUI is not available yet but the extensive configuration file should allow for enough flexibility and easy execution. Several output types like CSV, txt, html or hashlists should ease further processing of the findings depending on the investigators demands and needs. The hashlists work well to highlight the findings in Sleuthkit via the autopsy browser.
Further Roadmap
if my free time allows the following is already planned
- Multithreading to utilize multicore processors
- GUI for users not comfortable using the CLI
- implementation of further detection algorithms looking at other colorspaces like HSV
- implementation of a video decode library to scan frames of video streams
Some discussion of what has been implemented can be found here
Please take the time to test it and provide feedback - Thank you!
Looks interesting, though IMO the name's a bit naff. How does it cope with black/Asian skin tones?
Yes the name might not be the best - maybe you have another suggestion?
During my tests I also have a folder with ethnic diverse samples and due to the YCbCr detection algorithm it performs quite well.
http//
shows how the full 255^3 colorspace is indexed. For full discussion of this please goto
That's good, as a lot of skin detection analysis apps are weak at a range of skin colours. Re the name, I guess something that is self explanatory and less casual sounding… Skin Tone Detect? But that's a rather unfortunate acronym 😉
Well an unfortunate acronym can be lived with as => http//s-t-d.org/ - Linux has shown.
I will have a thought over the name… But you mentioned there are other skin detection applications - could you please provide more information about any free versions you know? A quick search on FF brought Nuix and X-Ways but these are not free.
Would be glad. Thank you!
Yes the name might not be the best - maybe you have another suggestion?
4skins?
lol lol lol
Well an unfortunate acronym can be lived with as => http//s-t-d.org/ - Linux has shown.
I will have a thought over the name… But you mentioned there are other skin detection applications - could you please provide more information about any free versions you know? A quick search on FF brought Nuix and X-Ways but these are not free.Would be glad. Thank you!
Those are the only ones I'm aware of, along with the add-on module for FTK 3.
Heh, and this is the reason Jamie should never be allowed to name software. Imagine standing up in court and saying you used that to identify the child images….
oops
Aw, c'mon, this is my best joke in years - surely I deserve a bit of credit (all it needs is a logo with a magnifying glass and even Jonathan would be happy!) I'm particularly proud of the "4" which I know all you 4ensicators love -)
It did make me chuckle, i'll give you that.
Have a gold star 