search

cbmi-group / FFM-Multi-Decoder-Network

The official code repository for the ECCV 2024 accepted paper "Representing Topological Self-Similarity Using Fractal Feature Maps for Accurate Segmentation of Tubular Structures".
9 stars 1 forks source link

RankWarning: Polyfit may be poorly conditioned #1

Open chinmay5 opened 2 months ago

chinmay5 commented 2 months ago

Thank you for releasing the code and congratulations on the ECCV acceptance.

I am trying to run the code using the ER dataset. However, the fractal_analysis.py file is giving me an error RankWarning: Polyfit may be poorly conditioned. Is this expected?

I am using

base_path = 'data/Fluorescence_Microscopy_Images_v2/cbmi_ER/train/images/'
# The step and window size as obtained from the paper
step_size = 1
window_size = 5
compute_FMM_Pool(base_path, window_size, step_size)

Best, Chinmay

ljmkgbkq commented 4 hours ago

Hello, I encountered the same problem as you when running fractal_analysis.py. Have you solved this problem? Looking forward to your reply~