Dahe Daily Yu video reporter Huang Ting
【Innovation】This is the only one in the world in Yinxu Anyang
Oracle bone script is an ancient Chinese script, also known as "qiwen" and "oracle bone buzhi", how can the hieroglyphs engraved on the bones 3,000 years ago be entered into the computer? The scattered research data and research forces of oracle bone and other ancient text research, and insufficient communication and coordination have been the pain points of oracle bone research for a long time. Building an Oracle database is particularly important.
On February 17, at Anyang Normal University, Liu Yongge, dean of the School of Computer and Information Engineering, introduced to the Henan reporter of Dahe Daily Yu Video Science and Technology that since 2000, he has led two teachers who study Oracle to carry out Oracle informatization research with Oracle input method as a starting point. In July 2018, with the approval of the Ministry of Education, Anyang Normal University established the Key Laboratory of Oracle Information Processing, which currently has four stable research directions of Oracle big data platform, Oracle bone recognition and glyph analysis, Oracle language computing, Oracle bone and Yinxu science and technology archaeology, and is the only science and engineering laboratory in the world that serves Oracle bone research.
"Because we want to do the big data of oracle bones, we must first collect all the paper materials on oracle bone research, at present, there are 7359 volumes of oracle bone documents collected in our data room, and I dare not say that all of them have collected 100% of the books on oracle bones at home and abroad, but there are also 99%." Liu Yongge said.
In 2019, the laboratory released the Oracle big data basic platform "Yinqi Wenyuan", which includes 152 oracle bone records, 23466 oracle bone images, 4487 oracle bone glyphs, 33368 oracle bone treatises, and provides free services to global oracle bone, archaeology, history, writing and other professional research experts.
From paper and ink transcription to Internet three-dimensional modeling, the ancient text of 3,000 years ago has instantly become flexible, not only ranking first in the "Oracle Bone Golden Text" data ranking of Peking University, but also attracting paleography enthusiasts, teachers and calligraphy enthusiasts at home and abroad to consult and use.
【Results】Artificial intelligence allows oracle bones to be accurately reproduced
The oracle bone text detection system based on deep neural network organizes and digitizes fragmented data and literature, and oracle bone researchers from the previous library to the current network database, using pinyin, Chinese characters, radicals to search, free reading of oracle bone images, interpretations, related literature, heavy films, conjugation library, glyph library information.
The Dahe News Yu video reporter observed that the "Yinqi Wenyuan" database has established a handwritten oracle bone character number set (HWOBC dataset), oracle bone text detection dataset, and oracle bone topology recognition dataset (OBC306 dataset), which use knowledge graph technology and diffusion network technology to associate all the relationships of oracle bones, words, texts, and diagrams, and then use computer logic to reason.
Liu Zhao, a professor at Fudan University and director of the Center for Excavated Documents and Paleography, said that since the discovery of oracle bones 120 years ago, about 160,000 oracle bone collections have spread around the world, and these scattered ancient information is the root vein of China's excellent traditional culture, and it is urgent to decipher and digitize it.
"In the past, oracle bone fragments were pieced together piece by piece by memory and by hand. After the data is digitized, it can be automatically conjugated using artificial intelligence image technology. At present, we have conjugated 37 groups, which are 37 groups that human experts have not yet conjugated. Liu Yongge said proudly.
In order to start a comprehensive survey and collection of global oracle bone data, the Oracle Key Laboratory of Information Processing has also used the latest scientific and technological means to develop Oracle bone three-dimensional modeling technology and equipment, obtain oracle bone rubbing, color photos, size, collection information, etc., and will carry out the first batch of Oracle 3D modeling experiments in the near future.
Liu Yongge introduced to reporters: "The original oracle bone graphics are from photos and rubbing, last year, the laboratory and Tencent jointly developed and created high-precision, high-speed, portable oracle bone three-dimensional modeling, adjusted and improved the existing core algorithm, proposed a set of point cloud fusion methods, completed 12 oracle bone three-dimensional models, now scan the QR code can be seen, to achieve the accurate reproduction of oracle bone collections." ”
【Breakthrough】The first definition of the Oracle information model
"This is the first time the lab has released a digital model of Oracle full information." On February 17, at Anyang Normal University, Liu Yongge told Dahe News Yu video reporter: "This is the first time in the world that a full-information data model has been proposed. Data fusion is carried out hierarchically, and high-fidelity interactive comparison display is realized. Our lab has discussed it many times, and the final name has not yet been decided. ”
On the analysis map shown by Liu Yongge, the reporter saw that the Oracle all-information data model developed by the laboratory is to classify and stratify information, similar to a mobile phone map, in a coordinate alignment manner, so that the oracle bone pieces are layered according to categories, and each audience can see different information according to different needs. These information levels include annotation, scientific images, three-dimensional data, rubbing, facsimile, basic data, etc. The traditional oracle bone bibliography is relatively single with pictures and text annotations.
Taking the shooting of oracle bones as an example, traditional cameras are limited by the light and the technical level of the photographers, making it difficult to shoot the fine lines on the oracle bones completely. In the all-information data model, using high-fidelity data shooting technology, the use of lighting transformation matrix can enlarge the photo dozens of times, not only can distinguish the oracle bone "pseudo-film", microtraces, trace elements, but also clearly see the oracle bone text ink that penetrates under the oracle bone and cannot be seen by the human eye.
The study of oracle bones is inseparable from archaeology, and the leap in oracle data information technology has also injected scientific and technological impetus into the archaeology of Yinxu. At the Xiaotungu Station in Anyang, the laboratory plans to conduct 3D modeling of 40 square kilometers of Yinxu to provide relevant knowledge and support for Oracle information processing.
The reporter learned that the laboratory has cooperated with Tencent to develop the "Great Oracle" mini program, which is expected to conduct internal testing in March, and plans to announce the project to the world in April. Liu Yongge said that this mini program mainly introduces and disseminates Oracle knowledge, and develops gameplay such as level breaking mini games and passing rewards. Use the communication efforts of WeChat users to let more public reach Oracle bone culture.
【Inheritance】There are heirs from the Flying Out Book Zhai generation
In the corridor of the Oracle Key Laboratory of Information Processing, there is a photo of the first Oracle Information Processing International Symposium in 2018, with 44 people in the photo. Liu Yongge said that this is already most of the world's experts in the study of Oracle information processing.
Oracle information processing, simply put, is to collect all the data of Oracle, and use technical means to develop and utilize data again. Because there are too few people studying, it once seemed that this discipline was very unpopular.
"According to the requirements of the Ministry of Education for key laboratories, it is very rare for Anyang Normal University to invite Dai Haiqiong, academician of the Chinese Academy of Sciences and professor of Tsinghua University, as the director of the academic committee. Song Zhenhao, a leading figure in oracle bone science, a member and researcher of the Chinese Academy of Social Sciences, donated his life's research results and data to the laboratory free of charge. Liu Yongge said emotionally that it is precisely because of the selfless dedication and guidance of scientific research predecessors that the "Yinqi Wenyuan" database platform can be accelerated and released to the world.
Liu Yongge said: "General Secretary Xi Jinping's congratulatory letter to the 120th anniversary of the discovery and research of Oracle bone bone points out the direction for our future work. The laboratory closely focuses on the major strategic needs of the country, takes service culture inheritance and innovation as the main line, implements the operation mechanism of "openness, mobility, alliance and competition", and strives to build a multidisciplinary collaborative innovation, international level Oracle big data platform and a first-class scientific research, technological innovation and talent training base in the field of Oracle information processing. ”
Since 2000, the research on oracle input method, 23 years of exploration, the laboratory will study the interdisciplinary integration of writing and cultural inheritance. At present, the members of the laboratory have undertaken 7 projects of the National Natural Science Foundation of China, 3 major commissioned projects of the National Social Science Foundation of China, 8 projects of the National Social Science Foundation of China, and more than 50 provincial, ministerial and prefectural and departmental projects, involving science, technology, humanities and art.
With the support of Liu Yongge, Anyang Normal University also opened a special course on Oracle Information Processing. More and more college students, from the beginning of their freshman year, have participated in Oracle informatization research and Oracle big data platform information collection and entry.
Liu Yongge said: "Using a computer to decipher oracle bones and using the reasoning of knowledge graphs can complete or predict the missing oracle bone semantics, so as to realize the automatic discovery and archaeological prediction of oracle bone knowledge." Our lab's ultimate dream is to decipher a new oracle bone script, and we are moving closer to that goal. ”
Source: Dahe News Yu Video Editor: Zhang Xuanzi