Changsong LIU
|
- Associate Professor
- Supervisor of Doctorate Candidates
- Supervisor of Master's Candidates
- Name (Simplified Chinese):Changsong LIU
- Name (English):Changsong LIU
- E-Mail:
- Date of Employment:1992-03-01
- School/Department:Tsinghua University
- Degree:Doctoral degree
- Professional Title:Associate Professor
- Status:Employed
- Alma Mater:清华大学
- Teacher College:DZGCX
Contact Information
- ZipCode：
- Fax：
- PostalAddress：
- OfficePhone：
- Email：
- Research Status
1. Tsinghua-Nokia Joint Research Program, “Handwriting and image recognition”, May. 2008-May. 2010, Project Leader.
The objective of the project is the development and optimization of the continuous Chinese multi-character HWR algorithm, development handwritten mathematical expression calculator, and some research on human face location and recognition. The handwritten mathematical expression calculator is selected and pre-installed on many new Nokia brand mobile phones.
2. Hi-tech Research and Development Program of China (National 863 Project), “Comprehensive platform for intelligent Chinese information processing” Jan. 2001-Dec. 2003, Principle Investigator.
The objective of the project is to improve Chinese OCR and handwritten character recognition performance and build some related applications. The result include an Asia language OCR engine which is licensed by Microsoft in there Office 2003 product, an online handwritten recognition engine which is used widely on many PDAs and mobile phones, and serval other systems.
3. National Science Funding, China, “Using style consistency information between characters to improve the character recognition system performance ”, Jan. 2008-Dec. 2010, Principle Investigator.
There are relations between the shape of one character with another in the same sentence or paragragh, We research on how to modelling these kind of information and use it improve character recognition.
4. Microsoft Cooperation, “Online handwritten character recognition for Tablet PC platform”, 2007, Principle Investigator.
The project produced a new full screen handwritten words and sentence recognition and result editing method and a demo system. The key problems resolved were using dictionary and language model to get good string recognition accuracy and invent a efficient way to edit the result.