Create a photo forensic tool using Python

We are working on a tool to help us improve our identity verification process. Our users would submit id documents like driving license to get verified on the platform. The tool should be able to detect the quality of submitted photo by the user like checking if the photo submitted by the user is blurry, has small size or low dpi, or there have glare on the photo. Other requirements are to detect if the image is manipulated, extract all the faces from the photo, perform face matching if the user submit a selfie with the id submitted (detect if the selfie matches the face on the id),extract all the text on the image using OCR. Here is the full list with some examples of how this can be done. 1) Detect photo manipulation (error level analysis (ELA)) [login to view URL] 2) Metadata extraction [login to view URL] [login to view URL] 3) Thumbnail consistency [login to view URL] 4) Detect glare on the photo [login to view URL] 5) Detect blurry images [login to view URL] 6) Extracting Face and Face matching [login to view URL] [login to view URL] 7) Extract all the texts from photo using OCR [login to view URL] [login to view URL] The code should be in Python and should be able to accept a single image (photo id), or two images (selfie and photo id), and return a json file with the results on each analysis mentioned above (a section for each). The performance should be reasonable as well to be able to use it in production. Python 3.x is preferred.