Many manga are produced through various processes such as line drawing creation, frame layout configuration, screen tone processing and phototypesetting. With the expansion of functions installed in the illustration tools and the support from them in recent years, even beginners without specialized skills have more opportunities to draw manga easily. However, it is difficult for beginners to draw a background on every frame in manga, and drawing them takes a lot of time. As one of the solution, it is an efficient drawing way to use line drawing materials created by another creator or created separately from characters for your manga. 

Examples of using the material (left) distributed in CLIP STIDIO PAINT for manga.
In the center frame, the material is inserted as it is, and in the right frame, it is traced again to match the creator’s style.

However, materials created assuming that it will be used by many creators places importance on versatility. So if you insert it as a background in manga as it is, Creator’s habits in drawing like a change in line thickness or fluctuation cannot be added to line drawing materials. Therefore, there are cases where materials not get used to hand-drawn objects. The use of materials reduces the time required to design the background and draw images according to the perspective, but it does not eliminate the labor required to adapt the materials to manga. The process of reflecting the characteristics of the drawing, such as the writing pressure and fluctuations that appear in the hand-drawn lines, on the material reduces the efficiency of drawing and become burdens to creators.

Examples of frames which have the background of a line drawing that applied our method to the input line drawing (left).
Each stroke texture and character used were drawn by same creator.

Therefore, we propose a method for semi automatically converting a vector line drawing of a building composed of line segments of uniform thickness into a line drawing like hand drawn one, and a system implementing the proposed method―LinDA (Line Drawing Artist). we quantitatively present what features each line segment of a line drawing has, and classify the line segments by the similarity of the features. we convert strokes in a batch according to the result of classification. This will create a line drawing that reflects the drawing characteristics of the creator based on the existing line drawing and stroke input.



NameAffiliationWeb site
Megumi NOMURAKeio University
Masanori NAKAYAMAKeio University



  1. Megumi Nomura, Masanori Nakayama, Issei Fujishiro: “LinDA: A semi-automatic system
    for generating cityscape background images―Converting atroke textures based on line segment features―
    ,”  The 6th autumn of  Federation of Imaging Societies of  Japan, Kyoto Institute of Technology, October 31―November 1, 2019 (in Japanese).
  2. Megumi Nomura, Masanori Nakayama, Issei Fujishiro: “LinDA: Feature extraction of line segments for semi-automatic generation of cartoon background images,” in Proceedings of the 81th National Convention of International Processing Society of Japan, Vol. 4, pp. 163―164,  Fukuoka University, Fukuoka-city, March 14―16, 2019 (in Japanese). Student Encouragement Award
  3. Megumi Nomura, Masanori Nakayama, Issei Fujishiro: “LinDA: Feature extraction and classification of line segments for semi-automatic generation of cartoon background images,” in The Technical Report of the Institute of Image Information and Television Engineers, Vol. 43, No.9, pp. 277―280,  Waseda University, Shinjuku, March 12, 2019 (in Japanese).


  1. Grant-in-Aid for Scientific Research (A): 17H00737 (2019-)
  2. Grant-in-Aid for Challenging Exploratory Research: 16K12459 (2018-2019)

Back to CA team page