HAND-DRAWN ELECTRICAL CIRCUIT RECOGNITION

Abstract

Hand-drawn sketch is a natural and direct way to express people's thought and meaning and is of common use in many different fields. Document image analysis is an active and challenging area of research in computer vision. Documents comprise of text and graphics. Machine recognition of hand-written text involves languages, mathematical symbols, digits, medical symbols etc. Machine recognition of hand-drawn graphical entities such as circuit diagrams, flow charts, tables, etc. will add another dimension to human computer interaction. In this work we propose a system of offline circuit recognition using digital image processing. The proposed model consists of all possible components of a diagram recognition system, such as segmentation, feature extraction, classification and redrawing and repositioning. we use topology based segmentation method to segment circuit sketch, and classify each component using the Fourier descriptors as feature vector for Support Vector Machine. An Accuracy rate of over 90% is achieved for each component.