Automated Recognition of Components in Hand Drawn Circuit Diagrams

Abstract

The project Automated Recognition of Components from Hand Drawn Circuit Diagrams developed to recognize circuit components from hand drawn circuit diagrams on paper. The hand drawn circuit diagram will be acquire as a scanned image to the application. Each component will be selected by user from circuit diagram to recognize. From which Histogram of Oriented Gradient (HOG descriptor) will extract features to match. A data set of trained features will also be maintained to match features from. Then Euclidean Distance will be calculated from trained features to get nearest neighbors. This way most repeating of nearest 5 components will be selected. Which will be labeled as a result in circuit diagram. User will also have a short description of recognized components.