- search engines
- AI image editing tools to enhance the quality of search results
- how to reverse face image search without photos
- quick face search on video
- protect yourself
🧵🧵🧵
#osint #socmint
When searching on Google, use not only Lens, but also the OLD image search (which gives very different results). You can access it using this tool:
A very good tool for searching for faces and images in general, which is often more effective than Google. It also recognises text in images and translates it.
This tool is paid, but sometimes it is more effective than Yandex and Google. Fortunately, you can find out if it will be useful for a particular photo for free (but the links are hidden).
The most important thing when gathering information about a person is to get the MAXIMUM number of photos of them. When one photo doesn't help you find anything, another may help you find everything
But if you have only one photo..... Do the others yourself!
🧵 7
Use AI image editing tools to change a person's age, eye colour, hair colour and skin colour. And then try to search for different image variants in the services described above.
Special attention should be paid to the direction of gaze (which can change facial proportions and affect search results). You can change it using this service (click "Look at bird"):
When a person takes selfies, their face looks different than when someone else takes their picture. Therefore, image mirroring can make a big difference in facial search results:
Very often the only images available are of the person in group photos, where the faces are blurred or too small. This is where AI services come in handy:
If the image cannot be reconstructed (or there is only an image stored in human memory), you can make a identikit of the person in and try to search by it ( works quite well)
Sometimes you have to cut out images of faces to search from the video. To quickly find images of all the characters that appear in a particular video (as well as different images of the same person), use this tool:
Is it possible to use neural networks to create a whole series of photos of one person for a fake social network profile ("sock puppet")?
I show by examples the possible way of solving the problem (this is not an easy way ).
🧵🧵🧵
In order to successfully pretend to be another person on the Internet, you need a series of photos in different places and different clothes. But it's a lot of work to make these (even with AI).
But often this leads to poor quality results. But for some situations it is quite good.
(2/7)🧵
It is better to extract text from a presentation so you can translate it later with more advanced translators. This can be done with the help of these services:
Ask a question and get an answer in the form of a list of research papers. Each one will have a text that refers to your question + the service will generate a summary of all the articles at once