A study of detecting line drawings from sketch image

Abstract

The objective of a study of detecting line drawings from sketch image is to study the detection of line drawings from sketches and turn the line drawing into a computer graphic which has 3 main steps. First, change the imported image to a gray-scale image. This process change the imported image property to gray-scale image which is easier to detect the border pattern. Second, smooth the image by using the Gaussian filter to reduce noise and unwanted streaks from the image. Finally, detect image border by using Canny edge detection which is a method for detecting image borders and compare the results of this algorithm with the Sobel edge detection method. With 50 sketches, we found that our process is able to reduce the noise well, gives smaller streaks and the resulting image is accurate with the details inside the image matching the original image.