The problem AI Verifica solves

We have built a scalable enterprise level solution that will help the computer identify if the image uploaded is acceptable in important documents, like PAN card and Voter IDs, this will reduce the need for manual labor and people chasing the government officers and bankers to change minor details sometimes leading to repeating the entire process multiple times.

We have multiple features implemented in our robust product.
Facial recognition and verification, the key features of our product, will be used to recognize and verify your facial images to enable extremely fast and secure signing in and verification process.

We also have features such as-
Face mask detection to enforce safety protocols to be followed in the pandemic.
Detection of blurry images, detection of sunglasses and detecting the number of people in an image. These features are useful for ensuring that proper images are being submitted by the user.

Better UI design Our product is built with the intention of making the user’s experience easier and secure.

Features :

1. Identity Authentication, used across different industries for ensuring safety and authenticity of the users.
2. Image Verification
3. Detects blurry images.
4. Displays the number of faces in an image
5. It detects if there are any obstructions in the image, like face masks, sunglasses in image, hands covering the face.
6. Overall acceptance score of photo is determined

Challenges we ran into

Challenges we ran into:

1. The extra feature of detection of photoshopped and original Picture which we tried implementing was giving only 50 % accuracy. So researching those tools took a lot of time.
2. Since we have built the image detection models from scratch, finding and cleaning the image datasets was a difficult task.
3. Incorporating the flask api and react front - end connection was a tumultuous task as we had never done this before.
4. Front-end side there was a phantom image phenomenon, debugging and figuring this out took a lot of time.