Loading...

Designing and Implementing an Enhanced Classification Algorithm in Image Processing

Baghery Daneshvar, Mohammad | 2013

2732 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: English
  3. Document No: 45195 (55)
  4. University: Sharif University of Technology, International Campus, Kish Island
  5. Department: Science and Engineering
  6. Advisor(s): Babaie-zadeh, Massoud; Ghorshi, Alireza
  7. Abstract:
  8. Statistical learning plays a key role in many areas of science [38]. An example of learning problems is image matching, image matching plays an important role in many aspects of computer vision.Computers can be used in intelligent tasks, which are followed by logical inference, for example, visual scenes (images or videos) or speech (audios). For humans visual system of such task are performed hundreds of times every day so easily sometimes without any awareness. In this thesis we focus on the image matching phase which is the first phase of the classification process. One of the popular image matching methods is Scale Invariant Feature Transform (SIFT) which our proposed method is based on it. The main idea behind our method is adding the following steps to SIFT:
    1.In the keypoint localization phase we added some steps to remove the excess keypoints.
    2.In feature description phase in addition to orientation histograms we added oriented patterns.
    3.In descriptor formation phase we decreased the size of the descriptors.
    By adding these changes to SIFT, we would have oriented patterns of keypoints. In addition, the number of keypoints have been reduced and the places of them would be selected more accurately, and the size of the descriptors have been reduced
  9. Keywords:
  10. Feature Extraction ; Image Matching ; Descriptor ; Oriented Pattern ; Image Processing

 Digital Object List

 Bookmark

  • Chapter 1 Introduction
    • 1.1 What is Image Matching
      • 1.1.1 Correlation Based Methods
      • 1.1.2 Feature Based Methods
    • 1.2 What is SIFT
    • 1.3 Why Implement SIFT
    • 1.4 Thesis Overview
  • Chapter 2 Background
    • 2.1 Feature Detection
      • 2.1.1 Guassian Derivative Based Detectors
      • 2.1.2 Other Technique Based Detector
    • 2.2 Feature Representation
      • 2.2.1 Gradient Distribution Based Descriptors
      • 2.2.2 Intensity Based Descriptors
      • 2.2.3 Derivative Based Descriptors
      • 2.2.4 Other Technique Based Descriptors
    • 2.3 SIFT
      • 2.3.1 Flow of Data in SIFT
      • 2.3.2 Scale Space Extrema Detection
      • 2.3.3 Keypoint Detection
      • 2.3.4 Orientation Assignment
      • 2.3.5 Keypoint Descriptor Generation
    • 2.4 Improvement on SIFT
    • 2.5 Summary
  • Chapter 3 Thesis Contribution
    • 3.1 An Improved SIFT Using Image Oriented Patterns (SIFTOP)
    • 3.2 Building Image Gaussian Pyramid
    • 3.3 Keypoint Extraction
      • 3.3.1 Exterma Detection
      • 3.3.2 Thresholding
      • 3.3.3 Eliminating Edge Response
      • 3.3.4 Localization
    • 3.4 Orientation Assignment
    • 3.5 Keypoint Descriptor
      • 3.5.1 Orientation Histogram
      • 3.5.2 Pattern Extraction
      • 3.5.3 Descriptor
      • 3.5.4 Training
    • 3.6 Object Recognition
    • 3.7 Block Diagram of the SIFTOP
    • 3.8 Summary
  • Chapter 4 Simulation Results
    • 4.1 Software Testing
    • 4.2 Method Testing
    • 4.3 Module Testing
    • 4.4 System Testing
    • 4.5 Summary
  • Chapter 5 Critical Evaluation
    • 5.1 Image Set
    • 5.2 Rotation Database creation
    • 5.3 Evaluation Criterion
    • 5.4 Experiment Results
      • 5.4.1 Image Rotation
      • 5.4.2 Scale Change
      • 5.4.3 Image Blurring
      • 5.4.4 Illumination Change
      • 5.4.5 Gaussian noise
    • 5.5 Demonstration System
    • 5.6 Summary
  • Chapter 6 Conclusion and Future work
...see more