As a comprehensive interdisciplinary laboratory, the Key Laboratory of Intelligent Press Media Technology (KLIPMT) was established at the end of 2016. KLIPMT focuses on knowledge mining and service, data management and operation, copyright protection technology. Aiming to meet the requirement of the press and publication industries, KLIPMT tracks the development of key technologies and their applications in related fields.
The laboratory currently has 38 faculties, including 7 professors, 7 associate professors or senior engineers. Having studied for more than 20 years in the press field, most faculties possess strong abilities in scientific research and innovation. Moreover, with an outstanding team adept in transferring research results into applications, KLIPMT can undertake major national research projects and participate in international competitions.
The current research in KLIPMT includes the following areas:
• Knowledge mining and service: focusing on distilling and mining knowledge resources from massive web data, improving open-domain knowledge base population, and designing algorithms to efficiently manage large-scale structured data, developing knowledge based tools or software to facilitate intelligent services, investigating novel and smart solutions for next-generation knowledge services, and exploring new technical standards for industries.
• Copyright protection: copyright protection technology and information hiding technology for a variety of media formats, service models, and terminals; improved effectiveness and versatility of copyright infringement tracking or authentication under the "Internet plus" application environment.
• Machine writing: constructing a large scale compositional semantic annotations for Mandarin Chinese, and studying natural language generation techniques, with particular interest in meaning-to-text models, as well as sentence compression and fusion based on deep learning.
• Intelligent document analysis and recognition: document layout analysis, automatic generation of compound document content (e.g., fixed-layout and stream), and semi-structure object retrieval.
• Character shape computing technology: quick generation of Chinese fonts based on stroke and radical assembling, intelligent CAD methods, deep learning based Chinese handwriting style representation and modeling, construction of large-scale Chinese font manifold.
• AR in the press: restoring three-dimensional objects from the contents in print media, and recognizing objects with AR methods.
• Comic content analysis: analyzing the cartoon page layout in image formats, understanding the semantic contents of comic.
• Development of interdisciplinary comprehensive application: document conversion and processing, copyright protection, and content service, focusing on the development of knowledge base and knowledge discovery systems with integrated multiple sources.