ABSTRACT

There is increasing interest in building systems that can automatically interpret hand-drawn sketches. However, many challenges remain in terms of recognition accuracy, robustness to different drawing styles, and ability to generalize across multiple domains. To address these challenges, we propose a new approach to sketched symbol recognition that focuses on the visual appearance of the symbols. This allows us to better handle the range of visual and stroke-level variations found in freehand drawings. We also present a new symbol classifier that is computationally efficient and invariant to rotation and local deformations. We show that our method exceeds state-of-the-art performance on all three domains we evaluated, including handwritten digits, PowerPoint shapes, and electrical circuit symbols. Electrical diagram is foundation of studies in electrical science. A circuit diagram conveys many information about the system. Behind any device there are plenty of electrical ingredients which perform their specific tasks, today all the electrical software tools failed to effectively convert the information automatically from a circuit image diagram to digital form. Hence electrical engineers should manually enter all information into computers, and this process takes time and brings errors with high probability. Moreover, when the diagram is hand drawn, the problem is more complicated for any electrical analysis. Thus, in this project we propose a new method using Artificial Neural Network (ANN) to make a machine that can directly read the electrical symbols from a hand drawn circuit image. The recognition process involves two steps: first step is feature extraction using shape based features, and the second one is a classification procedure using ANN through a back propagation algorithm.

Block Diagram :

A.   Hardware Requirement

Ø  System            :           Pentium IV 2.4 GHz.

Ø  Hard Disk         :           500 GB.

Ø  Ram                 :          4 GB

Ø  Any desktop / Laptop system with above configuration or higher level

Ø

B.   Software Requirements

Ø  Operating system                    :           Windows XP / 7

Ø  Coding Language                     :           Python

Ø  Interpreter                              :           Python 3.6

Ø  ML APIS                                  :           Sklearn, numpy, pandas, matplotlib,                                                                           tensorflow,keras