search

jianhuupenn / TESLA

Deciphering tumor ecosystems at super-resolution from spatial transcriptomics with TESLA
MIT License
42 stars 10 forks source link

Error in replicate tutorial #13

Closed Haz1y closed 7 months ago

Haz1y commented 7 months ago

Hi [Author's Name],

I hope this message finds you well. I've been trying to replicate the results you've achieved in your project, particularly regarding the neatness of the boundaries you've plotted. However, I seem to be encountering some difficulties in reproducing them accurately.

image image

Regardless of adjustments made to either the apertureSize or L2gradient parameters, the observed outcomes remain consistent. Could this discrepancy potentially be attributed to variations in OpenCV versions？

jianhuupenn commented 7 months ago

Thanks for your interest in TESLA! The contour in the tutorial was detected using:

cnt=tesla.cv2_detect_contour(img, apertureSize=5,L2gradient = True)

The image you show, which also looks good to me, seems generated from the scan_contour() function. The slight difference in the boundary region will not affect the results much.Feel free to proceed with your current contour