SamsungLabs / NeuralHaircut

Neural Haircut: Prior-Guided Strand-Based Hair Reconstruction. ICCV 2023
Other
516 stars 48 forks source link

Example Result #16

Open jiayuanlu opened 10 months ago

jiayuanlu commented 10 months ago

Hi. Thanks for your great work. Has someone run the example code (https://github.com/SamsungLabs/NeuralHaircut/blob/main/example) and met this problem? I followed the step to build my environment and I used v100 to run the example code. I ran 100000 steps in stage one and then started stage two. However one of the results in pred_hair_strands of stage two is as followed and I think it's vague: pred_hair_strands_99500 Does anyone know how to solve this problem?

Thanks.

Vanessik commented 8 months ago

Hi @jiayuanlu! In this work, we focused on reconstruction quality by using RGB loss only as an additional photometric signal. So, the results for renders could be blurry.

Btw, did you follow all postprocessing steps (./preprocess_custom_data/) after finishing stage 1 and before starting stage 2 (in case you use your checkpoints and geometry)?