Computer Vision: Edge detect project

02a507e3f7f2b7a00c9be6bd1f902dd1?s=47 Jamie Ly
October 09, 2010

Computer Vision: Edge detect project

Results of an edge detection project are discussed.

02a507e3f7f2b7a00c9be6bd1f902dd1?s=128

Jamie Ly

October 09, 2010
Tweet

Transcript

  1. CIS 581: Computer Vision Ly, Pant

  2. •Standard approach ◦ Local maxima ◦ Hysteresis •Different Thresholds for

    different images ◦ Threshold is a function of average gradient value and standard deviation of gradient value ◦ Threshold= mean + factor*std. deviation
  3. •2 steps ◦ Find all local maximas. ◦ Keep only

    those maximas which are above the threshold. •Two matrices ◦ LM_H = (Mag_q > Mag_r or q = r) and (Mag_q > Mag_p or q = p) ◦ LM_L = LM_H | Mag_q > thresh_mag
  4. •Mask_s = Mag_s > threshold_hys and LM_L •Mask_os = Mag_os

    > threshold_hys and LM_L •Hysteresis = LM_H and Mask_s and Mask_os
  5. •Matrix-based approach •Computation of theta using cart2pol ◦ Questions regarding

    unit vector •Attempt at recursive hysteresis
  6. Results

  7. Results

  8. Results

  9. Results

  10. Results

  11. •Understand the algorithm •Ask more questions early •Start early (did

    this!) •Try the way you know first. Optimize later. •Edges too thick!
  12. THANK YOU Questions?