Automation of Sign and Photo Detection Using AI and OCR

Every year, candidates for several DGT programmes, including CTS, CITS, and others, sit for entrance and yearly tests. A technique for automatically identifying photos and signatures in submitted documents is suggested by this Naner. By offering an automatic software tool that can reorder the inconsistent uploads that students made on the site during their We can assist the company in lowering the quantity of faulty admission cards distributed to trainees with low computer literacy during the exam registration procedure. [1] This will be advantageous for trainees as well as for other candidates who must upload documents with a signature and a photo. This will be useful for trainees as well as for other candidates who must upload documents with a signature and a photo. In addition to a trained neural network for categorising faces and signatures, Haar cascades are utilised to improve facial feature identification. Tensorflow and Pytorch were used for signature detection, and once the signature was converted into editable and searchable text, we used GAN (an AI tool) to convert the noise-filled image into a clear one. Last but not least, to produce a result with improved validity, a weighted average of both models is generated.