1 / 20

V. Alarc ón Aquino O. Starostenko R. Rosas Romero J. Rodríguez Asomoza O. J. Paz Luna

Mammographic image analysis for breast cancer detection using complex wavelet transforms and morphological operators. V. Alarc ón Aquino O. Starostenko R. Rosas Romero J. Rodríguez Asomoza O. J. Paz Luna K. Vázquez Muñoz L. Flores Pulido. Contents. Introduction Microcalcifications

zyta
Download Presentation

V. Alarc ón Aquino O. Starostenko R. Rosas Romero J. Rodríguez Asomoza O. J. Paz Luna

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript

  1. Mammographic image analysis for breast cancer detection using complex wavelet transforms and morphological operators

  2. V. Alarcón Aquino O. Starostenko R. Rosas Romero J. Rodríguez Asomoza O. J. Paz Luna K. Vázquez Muñoz L. Flores Pulido

  3. Contents • Introduction • Microcalcifications • Wavelet Transforms • Experimental Results • Conclusions SIGMAP 2009

  4. Introduction • A mammography exam is used to aid in the diagnosis of breast diseases in women • The early detection of breast cancer is difficult due that small tumors and microcalcifications are very similar to normal glandular tissue • So, wavelet transform is employed in eliminate noise in mammogram’s images SIGMAP 2009

  5. Introduction • A system based on fuzzy logic has been reported in (Cheng, 1998) • A mathematical morphology study is reported in (Zhao, 1993) • A two stages method for segmentation and detection of MC’s with matched filters (Strickland 1996) • Wang (1998) detect MCs using the decimated wavelet transform and a nonlinear treshold SIGMAP 2009

  6. Microcalcifications • The breast tissue study was performed in radiology including magnetic resonance image and nuclear medicine • Using both methods it helped to decide the best theraphy for each lesion • Unfortunately it is not possible improve the visualization of present elements • Digital mammographs is preferred SIGMAP 2009

  7. Microcalcifications • Breast microcalcifications are commonly discovered in the radiological story of asymptomatic women • These are deposits of calcium at the thickness of mamary tissue and are represented as little white dots • The first sign of cancerous process. SIGMAP 2009

  8. Mammography Image Analysis Society database SIGMAP2009

  9. Microcalcifications • MC’s are small deposits of calcium that appear as diminutive white dots in the mammogram • Due to microcalcifications’s size, the detection of: • non-homogeneus background of mammograms (breast glandular tissue) • noise detection of MC’s is difficult SIGMAP 2009

  10. Wavelet Transforms • Is a mathematical tool that provides building blocks with information in scale and time of a signal • The process of wavelet transform of a signal is called analysis • The inverse process to reconstruct the analyzed signal is called synthesis SIGMAP 2009

  11. Discrete Wavelet Transform • Is a time-scale representation of a digital signal, obtained with digital filtering techniques • The signal is passed trough several filters with cut-frequencies at different scales • The wavelet’s family is generated by a mother wavelet defined by: SIGMAP 2009

  12. Complex Wavelet Transform • Is used to avoid limitations of DWT and to obtain phase information • Real and imaginary coefficients are used to compute amplitude and phase information SIGMAP 2009

  13. Bank filter for 1D DT-CWT Analysis • The form of the conjugated filters of one-dimensional DT-CWT is defined for • Where: • is the set of filter • is the set • and correspond to low-pass and high-pass filter for real part • and are in the imaginary part • The synthesis bank filter is realized with the pairs and SIGMAP 2009

  14. Proposed Approach • The five steps that conforms the method to detect MC’s are: • Mammogram’s sub-band frequency decomposition • Mammogram’s noise reduction • Suppression of bands containing low-frequencies • Dilatation of high-frequency components • Mammogram’s reconstructionDetection of Microcalcifications SIGMAP 2009

  15. Experimental Results • Evaluation using the SWT and the Top-Hat Transformation • In the SWT case the fourth order Daubechies wavelet is used • The detection of MCs using the SWT is accomplished by setting low frequencies subbands to zero • Before the reconstruction of the image SIGMAP 2009

  16. Experimental Results Glandular tissue that contains a set of maligns MCs using DT-CWT SIGMAP 2009

  17. Experimental Results • Glandular tissue that contains a set of maligns MCs using SWT and Top Hat Transform SIGMAP 2009

  18. Experimental Results • SWT complexity is high O(n2) • DT-CWT O(2n) • Top-Hat transformation worst method to detect MCs • This is due to the fact that other tissues and breast’s glands are not filtered and appear together with MCs • Which are not significantly appreciated as in the cases of the two other simulated methods SIGMAP 2009

  19. Conclusions • SWT detects the MCs but other details are also observed as MCs • Inconvenient presented by the SWT computational complexity, O(n2) • Computational complexity of the DT-CWT is O(2n) only SIGMAP 2009

  20. THANKS! QUESTIONS?

More Related