Open AccessA Deep Learning Approach for Generating Mark-up Code from Sketch Images
Author(s) -
Bhavesh Lohana,
Muskan Tanna,
Gautam Pamnani,
Tanish Sahijwani,
Rohini Temkar
Publication year - 2022
Publication title -
international journal for research in applied science and engineering technology
Language(s) - English
Resource type - Journals
ISSN - 2321-9653
DOI - 10.22214/ijraset.2022.40675
Subject(s) - computer science , deep learning , workflow , sketch , code (set theory) , convolutional neural network , artificial intelligence , key (lock) , interface (matter) , process (computing) , encoder , programming language , database , operating system , algorithm , set (abstract data type) , bubble , maximum bubble pressure method
User Interface (UI) design is an important part of software development. Creating an intuitive and engaging user experience is a key goal for businesses of all sizes and is a process driven by rapid prototyping, design, and user testing cycles. It requires a significant amount of money and effort just to build a production-grade website. It's difficult to generate code from photos. The insight goal is to use modern Deep Learning algorithms to significantly simplify the design workflow and enable any business to quickly create and test web pages. The proposed Deep Learning model consists of a Convolutional Neural Network (CNN) encoder segment and a Gated Recurrent Network (GRU) decoder segment which is trained on a custom database of wireframe sketches and their corresponding code. The network will produce the HTML code, corresponding to the sketch image that is fed into the proposed model. Keywords: Computational Neural Network, Deep Learning, Machine Learning, Gated Recurrent Unit, Mark-Up Code Generation