Doodle a new way to communicate

Doodles are hand-drawn sketches and computers having the ability to understand our quick line drawings will allow for broader forms of expression and communication. In our project, we use machine learning techniques for efficiently and accurately recognize the label of hand-drawn sketches. The recog

Project Title

Doodle a new way to communicate

Project Area of Specialization

Artificial Intelligence

Project Summary

Doodles are hand-drawn sketches and computers having the ability to understand our quick line drawings will allow for broader forms of expression and communication. In our project, we use machine learning techniques for efficiently and accurately recognize the label of hand-drawn sketches. The recognized label then used for communication, searching or learning purpose by user.

Project Objectives

The core objective of this project is to enable computer to understand human free-hand sketches. The system will take hand-drawn sketch as input and analyze the sketch. The system will classify sketch and generate a label. The label is name of the class from which the drawn sketch belongs. Our aim is to improve the accuracy and efficiency by the current systems.

Improving the accuracy of hand-drawn sketches is also one of the main objective. Improving the accuracy will result in a system that might be used in real world.

Making an efficient system is also a major objective. Efficiency i.e. model size and training time is critical to allow deployment of the system in real-life application

Project Implementation Method

In development of our system, we are not using customary software development methods because in customary systems task can only be performed by writing a line of code for them and those systems cannot make predictions. 

We are using machine learning for developing our system. It is an application of Artificial Intelligence that provides systems the ability to automatically learn and improve from experience without being explicitly programmed. Machine learning focuses on the development of computer programs that can access data and use it learn for themselves.

The process of learning begins with observations or data, such as examples, direct experience, or instruction, in order to look for patterns in data and make better decisions in the future based on the examples that we provide. The primary aim is to allow the computers learn automatically without human intervention or assistance and adjust actions accordingly.

This is a research and development project. So, the development techniques may vary as progress is made in the project. Convolutional Neural Network(CNN) will be used in the project. A neural network will be used which based on CNN. A model will be generate in order to classify sketches accurately.

The system will be implemented by using Python 3.5 in the Anaconda environment. Libraries and frameworks such as OpenCV2, Tensorflow, Theano, Keras, Sci-kit Learn, Pandas, Matplotib, Numpy, OS, Tinkter will be used to assist in the development of the machine learning model and evaluate it. Supervised Machine learning with Convolutional Neural Network combined with image analyzer to develop the system. The proposed technique might change according to the project progress.

Benefits of the Project

 People travel to other countries for many reasons and they face problems when they are not familiar with the language of that country. Shopkeepers don’t understand what they are asking for and many other communication problems with the people of that country. So, by using this system they can easily communicate with the people of different language.

This system can also use for child learning. A child will draw a sketch and he/she will get the name of the drawn sketch. In this way, a child can learn name of things.

Dumb people find difficulty while communicating with normal people. They use sign language which is difficult for normal people to understand. By using this system, dumb people can easily express themselves to others by using sketches.

It is difficult for illiterate people to use keywords for searching but they can draw rough sketch of objects which they want to search. So, this system can help them using hand drawn sketches instead of keywords for searching.

Technical Details of Final Deliverable

Our goal is to build a new machine learning model that can analyze and classify sketches. This model will be based on Convolutional Neural Network and Support Vector Machine. This model deploy on server because machine learning models require high computational power therefore we will require GPUs for training and testing purpose of our model.

A mobile application will also be developed, which provide sketching environment to the user. The user will draw sketch on mobile application and submit it. The submitted sketch move to the server as input and its classification will perform on server. After classification, the label of the sketch will generate. The label will represent the name of the object which has been drawn by user. This label will send back to the application and the user will use it for communication, searching or learning purpose.

Final Deliverable of the Project

Software System

Type of Industry

IT

Technologies

Artificial Intelligence(AI)

Sustainable Development Goals

Quality Education

Required Resources

If you need this project, please contact me on contact@adikhanofficial.com