指导教师推荐意见(4)

Abstract

Skew detection is an important step of pre-processing in the Document

Image Retrieval System and OCR System, affecting the performance of the

systems. Skew detection will be applied in Intellectualized Document

Image Browse ，Intelligent Copycat and Intelligent Scanner.

Based on the applications and foregrounds, this thesis researches the

document image with complex layout and designs an algorithm for skew

detection named BBL. A lot of experiments show that this algorithm is

accurate, robust and running fast.

In chapter 1 of the thesis, it will introduce Document Image Retrieval

system and OCR system and explain the effect of skew detection in the

systems. Some typical algorithms for skew detection will be introduced，

analyzed and compared in chapter 2. Chapter 3 is the most important part

of the thesis. It will discuss BBL Algorithm in details. In chapter 4，

Several experiments to verify the algorithm will be introduced , analysis

and comparison of the experimental results will also be given in this

chapter. Lastly, the conclusions and prospect of the work will be given.

Key words：

Document Image， Skew Angle， Detection, Image Retrieval

------------------------------------------------------------------------------