Skeletons of characters provide vital information to support a variety of tasks,e.g.,optical character recogni-tion,image restoration,stroke segmentation and extraction,and style learning and transfer.However,automati...Skeletons of characters provide vital information to support a variety of tasks,e.g.,optical character recogni-tion,image restoration,stroke segmentation and extraction,and style learning and transfer.However,automatically skele-tonizing Chinese characters poses a steep computational challenge due to the large volume of Chinese characters and their versatile styles,for which traditional image analysis approaches are error-prone and fragile.Current deep learning based approach requires a heavy amount of manual labeling efforts,which imposes serious limitations on the precision,robust-ness,scalability and generalizability of an algorithm to solve a specific problem.To tackle the above challenge,this paper introduces a novel three-staged deep generative model developed as an image-to-image translation approach,which signifi-cantly reduces the model's demand for labeled training samples.The new model is built upon an improved G-net,an en-hanced X-net,and a newly proposed F-net.As compellingly demonstrated by comprehensive experimental results,the new model is able to iteratively extract skeletons of Chinese characters in versatile styles with a high quality,which noticeably outperforms two state-of-the-art peer deep learning methods and a classical thinning algorithm in terms of F-measure,Hausdorff distance,and average Hausdorff distance.展开更多
基金supported by the“2030 Major Project-New Generation Artificial Intelligence”funded by the Chinese Ministry of Science and Technology under Grant No.2020AAA0106302the National Natural Science Foundation of China under Grant No.6187615.
文摘Skeletons of characters provide vital information to support a variety of tasks,e.g.,optical character recogni-tion,image restoration,stroke segmentation and extraction,and style learning and transfer.However,automatically skele-tonizing Chinese characters poses a steep computational challenge due to the large volume of Chinese characters and their versatile styles,for which traditional image analysis approaches are error-prone and fragile.Current deep learning based approach requires a heavy amount of manual labeling efforts,which imposes serious limitations on the precision,robust-ness,scalability and generalizability of an algorithm to solve a specific problem.To tackle the above challenge,this paper introduces a novel three-staged deep generative model developed as an image-to-image translation approach,which signifi-cantly reduces the model's demand for labeled training samples.The new model is built upon an improved G-net,an en-hanced X-net,and a newly proposed F-net.As compellingly demonstrated by comprehensive experimental results,the new model is able to iteratively extract skeletons of Chinese characters in versatile styles with a high quality,which noticeably outperforms two state-of-the-art peer deep learning methods and a classical thinning algorithm in terms of F-measure,Hausdorff distance,and average Hausdorff distance.
