hzuhdi / aware

Aware is DeepFake detection solution built on top of different existing pre-trained open source models.

Aware 👀

Aware is DeepFake detection solution 🌎 built on top of different existing pre-trained open source models

aware's Issues

As a user, I want my video to be validated for deepware by Aware Api

With the api endpoint created for receiving video, the api to be able to validate/detect the video for deepware using a machine learning model. Currently the machine learning model is in place, and can be integrated with the api.
For the integration, this investigation has been done in order to fulfil the requirements of what is needed for the API (.NET 6) to communicate with Python script (Machine Learning model language).
Approach :

• IronPython
• .NET

Acceptance Criteria

• Connect the machine learning model with the api
• The received video can be validated for deepware elements using the machine learning model

As a user, I want to be able to send a video to Aware Api

With the api created, the api has to have an endpoint to be able to receive video file.
There are certain video formats that can be accepted for now.
Once the video is received, it needs to be removed after processing/validating it for deepware elements.
Keep in mind the security considerations : https://docs.microsoft.com/en-us/aspnet/core/mvc/models/file-uploads?view=aspnetcore-6.0

Acceptance Criteria

• API should be able to accept this video formats; avi, mp4, ???
• Video should be removed after deepware detection/validation

Create deployment and dockerfile for Frontend

As a User I want to be able to upload my video to test if it is deepfake or not

Acceptance Criteria:

• The video file formats are either .mov or .avi
• The maximum file size of the video is 25mb
• A notification is shown after the file is uploaded that indicates whether the video is deepfake or not

As a User I want to be able to upload my Audio and detect if it is deepfake or not

Acceptance Criteria:

• The audio format is .wav
• The maximum audio file is 10mb
• A notification is shown that indicates whether the audio is deepfake or not

As a user I want to be able to upload my Text file so I will be able to detect if it is deepfake or not

Acceptance Criteria:

• The text file is in .txt format
• The maximum size of the text is 10mb
• A notification will be shown that indicates whether the text is deepfake or not

As a User I want to be able to see the leaderboards of user for each criteria (text, video, image, audio) who've uploaded correct information, so that it can motivate other users to share correct information

Acceptance Criteria:

• The content sharer can see the leaderboard of this month
• The content sharer can see the leaderboard of this year
• The content sharer can see his/her rank this month
• The content sharer can see his/her rank this year

As a User I want to be able to upload my image so it will detect if it is deepfake or not

Acceptance Criteria:

• The image input is in jpg or png format
• The max of the image file size is 5mb
• A notification will be shown that indicates whether the image is deepfake or not

As a User I want to be able to register to the Aware App and Login to App so I can continue using Aware App

No need to use Auth 0

• Just create dummy user

As a developer, I want to be able to integrate deepfake facial recognition in video with the Aware Api

This story covers how developers can integrate deepfake facial recognition in video with the Aware Api.

Acceptance Criteria

• Find out which ML Model/ tech related to the model, that can be integrated with Aware API

As a user, I want to increase my reputation point

Acceptance Criteria:

• The point is increased when user shares a real information
• The point does not change when user shares a deepfaked information

Create deployment and dockerfile for Backend

As a User I want to be able to unregister in the Aware App so that Aware App complied to GDPR

Acceptance Criteria:

• Users can delete their account
• Users' data footprint is completely deleted from database
• Users can register again with the same email

As a Content Sharer I want to see my reputation point

Acceptance Criteria:

• The content sharer can see his/her point

As a User I want to be able to see all my uploaded resources in my profile so I know all resources I have uploaded

Acceptance Criteria:

• User can see a list of uploaded resources (text, image, audio, or video)
• User can filter based on the resource type (text, image, audio, or video)

As a user, I want my video to be detected whether or not it is deepfake.

With this inference model, the Aware system can detect deepfake in a single video or collection of videos. This inference uses the pre-trained model from Deepware that can be run in the script that will be called by the backend of Aware. For the requirement, it will need cuda. There should be also an alternative if the requirement cannot be fulfilled.

Acceptance criteria:

• This deepfake inference can be called by the backend system of Aware and give the result of the prediction.

As a developer, I want to have a simple API for Aware solution

To begin our step in creating a solution that could tackle the deepfake issue, an api is needed to bridge our consumers with the Machine learning models in order to share our service later on.

Acceptance Criteria

• Create a simple API
• This api can be accessed in live environment