Question about downsampled mesh performance

mks0601 / I2L-MeshNet_RELEASE

Official PyTorch implementation of "I2L-MeshNet: Image-to-Lixel Prediction Network for Accurate 3D Human Pose and Mesh Estimation from a Single RGB Image", ECCV 2020

MIT License

708 stars 130 forks source link

Question about downsampled mesh performance #117

Closed MooreManor closed 1 year ago

MooreManor commented 1 year ago

@mks0601

Do you tried downsampling the mesh to a smaller one (e.g. maybe 1000 vertices) and then train the network? Maybe this can reduce the redundancy and train faster. How about the performance?

mks0601 commented 1 year ago

I haven't tried, but that could be useful for efficiency.