sigseg5/nometa-tg

This is a Python Telegram bot that provides an opportunity to automatically delete all metadata from photo and apply Fawkes tool to hide you face from face recognition apps in social networks etc.

nometa-tg

This is a Python Telegram bot that provides an opportunity to automatically delete all metadata from photo and apply Fawkes tool to hide you face from face recognition apps in social networks etc.

Info

By default, this bot automatically remove all metadata from photos and apply Fawkes at a minimal mode. You can specify Fawkes settings (FAWKES_MODE) in .env file.
At this release you can send a photo as photo or document.You will get message if tool can't find any faces.
In future release will be implemented settings manager, and may be video processing.
However, you can find full information about Fawkes tool at GitHub, and this academic paper

Requirements

• Telegram account
• (Recommended) Installed Docker

Usage

Choosing FAWKES_MODE

FAWKES_MODE the tradeoff between privacy and perturbation size. Select from low, mid, high. The higher the mode is, the more perturbation will add to the image and provide stronger protection.
Feel free to play with mods and choose that mostly suits you
In the image below you can see an example of how the Fawkes tool works (photo from the Fawkes GitHub page, deprecated).

Run in Docker

• Add TOKEN to .env file. You can get token from BotFather at Telegram
• Run docker build --tag nometa-tg . for build image
• Run docker run --env-file .env nometa-tg to start or sudo docker run -d --restart unless-stopped --env-file .env nometa-tg if you want automatically starts container after server reboot Note: Maybe you should add sudo before all Docker-assigned commands

Run locally

Theoretically should work with Python 3.7, can't test it yet. I can't guaranty that this will be works, please use Docker

• Create images and documents folder in project folder
• Add TOKEN and FAWKES_MODE environment variables to your IDE config e.g. PyCharm. You can get token from BotFather at Telegram. For first run you can set FAWKES_MODE=min
  E.g. ENV_VAR: PYTHONUNBUFFERED=1;TOKEN=1234567890:xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx;FAWKES_MODE=min
• Run pip install -r requirements.txt to install dependencies
• Run bot.py to start

Docker control

Stop container

• Run docker stats to view all running containers
• Run docker kill <container_id>

Remove image

• Run docker images to view list of all images
• Run docker rmi -f <image_id>

Contribute

Any ideas or trouble? Please open issue or pull request

Communicate

Feel free to write me at Telegram

nometa-tg's Issues

Keras libraries problems

After I've built the docker image and run it, I've discovered some problems with keras library which is required by fawkes. I'm not sure, but maybe it could be fixed by specifying specific keras package version in reuirements.txt.

I've fixed it by entering the container (run docker exec -it $CONTAINER_NAME bash where $CONTAINER_NAME should be replaced with the name of container name which you can see by running docker ps) and changing some files:

• /usr/local/lib/python3.7/site-packages/fawkes/utils.py change import line from from keras.preprocessing import image to import keras.utils as image at the beginning and in resize function definition.
• /usr/local/lib/python3.7/site-packages/fawkes/differentiator.py change tf.keras.optimizers to tf.keras.optimizers.legacy

Then it start to work.

Here are some error:

Traceback (most recent call last):
  File "/usr/local/bin/fawkes", line 8, in <module>
    sys.exit(main())
  File "/usr/local/lib/python3.7/site-packages/fawkes/protection.py", line 201, in main
    separate_target=args.separate_target, debug=args.debug, no_align=args.no_align)
  File "/usr/local/lib/python3.7/site-packages/fawkes/protection.py", line 92, in run_protection
    image_paths, loaded_images = filter_image_paths(image_paths)
  File "/usr/local/lib/python3.7/site-packages/fawkes/utils.py", line 119, in filter_image_paths
    img = load_image(p)
  File "/usr/local/lib/python3.7/site-packages/fawkes/utils.py", line 109, in load_image
    image_array = image.img_to_array(img)
AttributeError: module 'keras.preprocessing.image' has no attribute 'img_to_array'

Traceback (most recent call last):
  File "/usr/local/bin/fawkes", line 8, in <module>
    sys.exit(main())
  File "/usr/local/lib/python3.7/site-packages/fawkes/protection.py", line 201, in main
    separate_target=args.separate_target, debug=args.debug, no_align=args.no_align)
  File "/usr/local/lib/python3.7/site-packages/fawkes/protection.py", line 128, in run_protection
    protected_images = generate_cloak_images(self.protector, original_images)
  File "/usr/local/lib/python3.7/site-packages/fawkes/protection.py", line 30, in generate_cloak_images
    cloaked_image_X = protector.compute(image_X, target_emb)
  File "/usr/local/lib/python3.7/site-packages/fawkes/differentiator.py", line 169, in compute
    target_imgs[idx:idx + self.batch_size] if target_imgs is not None else None)
  File "/usr/local/lib/python3.7/site-packages/fawkes/differentiator.py", line 248, in compute_batch
    optimizer.apply_gradients(zip(grad, [self.modifier]))
  File "/usr/local/lib/python3.7/site-packages/keras/optimizers/optimizer_experimental/optimizer.py", line 1140, in apply_gradients
    return super().apply_gradients(grads_and_vars, name=name)
  File "/usr/local/lib/python3.7/site-packages/keras/optimizers/optimizer_experimental/optimizer.py", line 634, in apply_gradients
    iteration = self._internal_apply_gradients(grads_and_vars)
  File "/usr/local/lib/python3.7/site-packages/keras/optimizers/optimizer_experimental/optimizer.py", line 1169, in _internal_apply_gradients
    grads_and_vars,
  File "/usr/local/lib/python3.7/site-packages/tensorflow/python/distribute/merge_call_interim.py", line 51, in maybe_merge_call
    return fn(strategy, *args, **kwargs)
  File "/usr/local/lib/python3.7/site-packages/keras/optimizers/optimizer_experimental/optimizer.py", line 1217, in _distributed_apply_gradients_fn
    var, apply_grad_to_update_var, args=(grad,), group=False
  File "/usr/local/lib/python3.7/site-packages/tensorflow/python/distribute/distribute_lib.py", line 2637, in update
    return self._update(var, fn, args, kwargs, group)
  File "/usr/local/lib/python3.7/site-packages/tensorflow/python/distribute/distribute_lib.py", line 3710, in _update
    return self._update_non_slot(var, fn, (var,) + tuple(args), kwargs, group)
  File "/usr/local/lib/python3.7/site-packages/tensorflow/python/distribute/distribute_lib.py", line 3716, in _update_non_slot
    result = fn(*args, **kwargs)
  File "/usr/local/lib/python3.7/site-packages/tensorflow/python/autograph/impl/api.py", line 595, in wrapper
    return func(*args, **kwargs)
  File "/usr/local/lib/python3.7/site-packages/keras/optimizers/optimizer_experimental/optimizer.py", line 1213, in apply_grad_to_update_var
    return self._update_step(grad, var)
  File "/usr/local/lib/python3.7/site-packages/keras/optimizers/optimizer_experimental/optimizer.py", line 217, in _update_step
    f"The optimizer cannot recognize variable {variable.name}. "
KeyError: 'The optimizer cannot recognize variable Variable:0. This usually means you are trying to call the optimizer to update different parts of the model separately. Please call `optimizer.build(variables)` with the full list of trainable variables before the training loop or use legacy optimizer `tf.keras.optimizers.legacy.{self.__class__.__name__}.'