While designing any product, Designers make a hand-drawn wireframe of the product and show them to the higher officials for review and generally, they can't able to related those hand-drawn wireframes to the product. So are finding a way which can solve this problem, basically, we are restricting this application to web pages {otherwise it would be a very large porject}. Basically, Automated Frontend Designer can turn a hand-drawn sketch into HTML code and instantly generate a webpage from the sketch at real-time. So designers don't have to depend upon the developers and prototyping would be easier than ever.
Since we have made it from scratch so we were not able to find a proper dataset to train our object detection model, so we decided to train the model using only 100 images. Since the dataset was too small so our model is not giving accurate results.
Also, Tesseract OCR is not much accurate on handwritten text, that is also the main concern. But finally after all these problems, we were able to finish the prototype and it is working really good.
Technologies used
Discussion