DEMO.md

Quick Demo

We provide demo codes for end-to-end inference here.

Our inference codes will iterate all images in a given folder, and generate the results.

Human Body Reconstruction

This demo runs 3D human mesh reconstruction from a single image.

Our codes require the input images that are already cropped with the person centered in the image. The input images should have the size of 224x224. To run the demo, please place your test images under ./samples/human-body, and then run the following script.

python ./metro/tools/end2end_inference_bodymesh.py 
       --resume_checkpoint ./models/metro_release/metro_3dpw_state_dict.bin
       --image_file_or_path ./samples/human-body

After running, it will generate the results in the folder ./samples/human-body

Hand Reconstruction

This demo runs 3D hand reconstruction from a single image.

You may want to provide the images that are already cropped with the right-hand centered in the image. The input images should have the size of 224x224. Please place the images under ./samples/hand, and run the following script.

python ./metro/tools/end2end_inference_handmesh.py 
       --resume_checkpoint ./models/metro_release/metro_hand_state_dict.bin
       --image_file_or_path ./samples/hand

After running, it will outputs the results in the folder ./samples/hand

Limitations

• This demo doesn't perform human/hand detection. Our model requires a centered target in the image.
• As METRO is a data-driven approach, it may not perform well if the test samples are very different from the training data.
• METRO is a mesh-specific approach. For example, our hand model is trained only on right-hand data, and therefore it doesn't perform well on the left-hand images. How to develop a unified model for different 3D objects is an interesting future work.