Design implementation of a sketch isolation algorithm: A computer vision-based approach

• Main Authors: Lindo, Delfin Enrique G., Cotoco, Ezekiel Karl A., Baldovino, Renann G., Bugtai, Nilo T.
• Format: text
• Published: Animo Repository 2017
• Subjects: Image processing—Digital techniques; Computer vision; Manufacturing
• Online Access: https://animorepository.dlsu.edu.ph/faculty_research/1386; https://animorepository.dlsu.edu.ph/context/faculty_research/article/2385/type/native/viewcontent
• Institution: De La Salle University

The process of digitizing sketches requires the intermediary process of scanning an actual paper. Through a computer-based program presented in this paper, the sketch, which is the foreground, is digitally extracted and isolated from the background creating a similar mask for a digital image. Through the sketch isolation process, prior sketches are excluded from the mask thus allowing individual editing of the most recent sketch. The gradient of the paper from the camera does not show in the final image, but instead shows a clean white background. The program does the process instantaneously allowing live sketch digitizing. It is suggested that the application be improved through the use of neural networks. The resolution of the final image was only as good as the input image, but it can be improved by converting it to a vector based graphic format the sketch will be easily scalable without losing detail through pixelation. © 2017 IEEE.