Dissertation > Excellent graduate degree dissertation topics show

Computer-Aided Detection Algorithm of Lung Nodules Based on Hessian Matrix

Author: ZhangJuan
Tutor: WangChangYuan；LuZuo
School: Taishan Medical College
Course: Medical Imaging and Nuclear Medicine
Keywords: Computer-aided diagnosis Computer-aided detection Hessian matrix Pulmonary nodules
CLC: TP391.41
Type: Master's thesis
Year: 2009
Downloads: 97
Quote: 3
Read: Download Dissertation

Abstract


Objective To study the Hessian matrix - based multi-scale enhancement algorithm of the enhanced role of the round or oval suspected nodule area for computer-aided detection of lung nodules help diagnose doctors beginning of the seizure of the pulmonary nodules , alleviate medical burden and improve the rate of detection of lung nodules . Materials and Methods This study selected 44 lung CT images , a total of 90 nodules of varying sizes as the experimental sample data , image size of 512 × 512 pixels , 16 bit allocation , store 12 -bit , 4096 grayscale level , 5 mm or 7 mm thick , the pixel size of 0.74 mm × 0.74 mm , tube voltage the 140kVp, tube current 300mA, images are stored as DICOM standard format . In Windows XP, VC 6.0 programming environment directly read CT images in DICOM format , and then the image is preprocessed by threshold segmentation and region growing method to extract the full lung parenchyma to eliminate redundant I information and the purpose of noise reduction of image information , the use of the Gaussian function to simulate lung nodules , construct the Hessian matrix , using a different scale parameters to achieve enhanced nodules of different sizes , to extract the region of interest of the suspected nodule , then CAD and no CAD clinical doctor's diagnosis by ROC curves were evaluated to calculate the area under the ROC curve . The results of this study to achieve the the multiscale Hessian matrix enhancement algorithm enhancements round class round lung nodules suspected area , the clinician diagnosis of CAD and non- CAD ROC analysis , The results show that there are CAD diagnostician the nodule detection ability (AZ = 0.95) higher than those without CAD (AZ = 0.88). Conclusion Hessian matrix - based multi-scale enhancement algorithm can effectively improve the round or oval lung nodule detection rate , but a higher false positive . The algorithm is practical and effective , the research and development based on the algorithm of the software system interface friendly and easy to use .

Related Dissertations

  1. The Research and Implementation of Lgorithms on Microcalcification Detection,R816.4
  2. Research on Diagnosis Methods of Breast Masses Based on Reference Images,TP391.41
  3. Research on Fast Corner Extraction Algorithm in Camera Calibration,TP391.41
  4. Research on Tracking Algorithm of Vessel Centerline in Coronary Angiography,TP391.41
  5. A Detection Algorithm Based on Morphology and Gray Entropy for Pulmonary Nodules,TP391.41
  6. Respiratory Gated 18F-FDG PET/CT Imaging Clinical Application Study of Pulmonary Nodules,R817.4
  7. Research of Microcalcification Detection and Computer-Aided Diagnosis Techniques Based on Scale Space Filtering,TP391.41
  8. The Application Research on Computer Aided Diagnosis of Brain Tumor Using Change Detection Techniques,TP391.5
  9. The Research of Key Techniques for Auto Recognition Tuberculosis Cells,TP391.4
  10. Planning system of computer-aided detection,F270.7
  11. Wavelet-based diagnosis of breast microcalcifications Algorithm,TP391.41
  12. Study on Recognition and Analysis of SPN Based on CT Images,TP391.41
  13. Pulse diagnosis , tongue secondary treatment system database creation,R241
  14. Research on Solitary Lung Nodule Detection of Low-dose CT Images,TP391.41
  15. The Clinical Value Study of Application of Video-assisted Thoracoscopic Surgery Diagnosis and Treatment of Solitary Pulmonary Nodule,R563.9
  16. Diagnosis of solitary pulmonary nodules clinical analysis,R655.3
  17. Based on the two-dimensional feature vector brain CT image registration and segmentation,TP391.41
  18. The Differential Diagnosis Value of 64-slice Spiral CT Perfusion Imaging in Solitary Pulmonary Nodules,R816.4
  19. Histogram analysis of solitary pulmonary nodules,R816.4
  20. Evaluate the Influence of Medical Monochrome Displays with Different Resolutions for the Interpretation of DR Chest Images,R816.4

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