Dissertation > Excellent graduate degree dissertation topics show

Research and Implementation on Text Recognition in Image with Complex Scenes Using Local Features

Author: WangHuiJing
Tutor: GuanHaiBing；ChenKai
School: Shanghai Jiaotong University
Course: Communication and Information System
Keywords: Local features Complex background Image processing Text Recognition
CLC: TP391.41
Type: Master's thesis
Year: 2010
Downloads: 206
Quote: 0
Read: Download Dissertation

Abstract


The image text recognition is Hot Spots and important issue of digital image processing and computer vision field. Compared to scanned images, complex background and natural captured images exist: 1, text fonts and stroke thickness; 2, the text characters arranged in a variety of layout; 3, complex and diverse background colors and textures; 4, the camera angle and carrier distortions caused by the character geometry deformation; 5, luminosity uneven noise result in low-resolution images, and other characteristics. The above features make for a complicated background image text recognition and natural shooting a difficult and challenging object recognition, rather than a simple optical character recognition (OCR). OCR on the text structure specification requirements and limitations on the input image, making the recognition framework based on OCR technology there is a big limitation. Although it is possible by improving the pre-text positioning and pre-processing sectors to provide OCR structured and standardized better input to obtain a certain increase of the recognition rate, but these aspects of the optimization is difficult and limited. Different from the frame of discernment based on OCR technology, this paper proposes a text recognition framework based on local feature. The framework of the principles and techniques of image retrieval, by building the template character image library, local image feature matching images with complex background text recognition. Different applications and processing means for local features, the Bag-of-Words Model-based recognition system based on the Point-to-Point Matching. Compared with OCR technology-based framework: 1, eliminating regional enhancement, binarization, layer analysis, geometry owned by a series of complex pre-processing aspects; 2, through the use of geometric and photometric invariance of local features and the introduction of targeted voting algorithm and geometric consistency verification, to overcome the limitations of OCR for text rotation, irregular arrangement, uneven image resolution, perspective transformation and distortions conditions identify; 3, by constructing multi-lingual and multi-font template word image library that implements the transparency and robustness of the recognition languages ??and fonts. This article focuses on Chinese, Japanese, Korean, English, Arabic, based on the single-character, multi-character, naturally captured images on a large number of comparative experiments, the results show recognition framework based on local features technology in processing power and recognition accuracy have good performance, more suitable for complex background and nature photography image text recognition.

Related Dissertations

  1. Research on Temperature Measurement Technique Based on CCD Image Sensor,TH811
  2. Research on Basic Algorithms of Digital Image Processing and Implementation with FPGA,TP391.41
  3. Research of Images Enhancing Algorithms on Fog or Backlighting Conditions and Implementation with Hardwares,TP391.41
  4. Research on the Image Real-Time Acquisition, Storage and Image Processing System,TP391.41
  5. The Research of Moving Object Tracking System Based on Embeded Image Process Unit,TP391.41
  6. Research on Visual Servo System of Mechanical ARM,TP242.6
  7. Application Research of Digital Image Processing on Container Inspection,TP274.4
  8. Three-Dimensional Numerical Modelling for the Head and the Electric Field Analysis of DBS,R742.5
  9. Application of Semi-Structure in Leisure Dress,TS941.2
  10. Research on Nondestructive Detection Technology for External Qualities of Papayas Based-on Vision,S667.9
  11. Research on Detecting Optical Fiber Geometric Parameter Based on Machine Vision,TN253
  12. Research on Identification System of Cashmere and Wool Fiber,TS101.921
  13. Design of Garlic Planting Sorting Machine Based on Image Processing,S223.2
  14. Study on Detection and Grading of ’Jiro’ Persimmon’s External Quality Based on Computer Vision,S665.2
  15. Flow Behavior Description of Two-Phase Flow Based on Image Processing Techniques,TP391.41
  16. Active faults based radar image information extraction method applied research and demonstration,P542.3
  17. Research on License Plate Location Algorithm of License Plate Recognition System,TP391.41
  18. The Designing of HD Vehicle Detection System Based on Fpga,TP391.41
  19. System Designing to Analysis the Real-time Density of Group-control Elevator Passengers Flow,TP391.41
  20. Research on Face Recognition Methods Based on Flexible Neural Tree,TP391.41
  21. Research on Automatic Reading System for Digital Meters,TP391.41

CLC: > Industrial Technology > Automation technology,computer technology > Computing technology,computer technology > Computer applications > Information processing (information processing) > Pattern Recognition and devices > Image recognition device
© 2012 www.DissertationTopic.Net  Mobile