KNN based Hand Drawn Electrical Circuit Recognition

Abstract

Electrical diagram is foundation in electrical science. A circuit diagram pass on numerous data about the framework. Behind any gadget there are a lot of electrical fixings which play out their particular tasks, today all the electrical programming instruments neglected to successfully change over the data naturally from a circuit picture diagram to computerized shape. Henceforth electrical designers ought to physically enter all data into PCs, and this procedure requires some investment and carry mistakes with high likelihood. Besides, when the diagram is hand drawn, the issue is more convoluted for any electrical investigation. Accordingly, in this paper we propose another strategy utilizing K-Nearest Neighbour (KNN) to make a machine that can straightforwardly read the electrical images from a hand drawn circuit picture.

Sketic: a machine learning-based digital circuit recognition platform

“…This improves the scalability of Sketic by allowing it to support new logic blocks, inputs, and outputs seamlessly. More importantly and as will be shown in the results section, the use of DNN and CNN machine learning algorithms causes the recognition accuracy of Sketic to reach 99%, which is higher than what was reported in previous studies [13,17]…”

“…The research efforts in this domain are not limited to digital circuits but they have been extended to recognize electrical circuits [15][16][17]. The authors in previous studies [15,16] applied image segmentation and feature extraction of electrical circuit elements (e.g., resistor and capacitor).…”

“…Dewangan and Dhole [17] used the shape area, major axis length, minor axis length, centroid, orientation, eccentricity, and convex area to generate the feature vector and then applied the K-nearest neighbor (KNN) algorithm to classify the circuit elements into their correct class.…”

“…The SVM, DT, and KNN machine learning algorithms used in the previously proposed techniques require careful and accurate feature selection and extraction in order to achieve acceptable recognition results. However, manually extracting the features as done previously [13][14][15][16][17] requires deep analysis of target elements. In addition, for each new digital block or circuit element the feature extraction procedure might be repeated with new features required to resolve conflicts and similarities.…”

Offline hand-drawn circuit component recognition using texture and shape-based features

“…Dewangan and Dhole [12] use a strategy utilizing K-nearest neighbor (KNN) classifier to build a system that directly reads the electrical circuit components from a hand-drawn circuit images. Different shape and geometrical features like area, length of major axis and minor axis, centroid position and orientation of the circuit components are considered.…”