---

This ReadMe provides instructions for using the BEDLAM2 (SMPL-X) dataset in training and evaluating CameraHMR. For the SMPL version, please refer to the main ReadMe.md.

Installation

Create a conda environment and install all the requirements.

conda create -n camerahmr python=3.10
conda activate camerahmr
pip install torch==2.0.0 torchvision==0.15.1 torchaudio==2.0.1 --index-url https://download.pytorch.org/whl/cu118
pip install -r requirements.txt

Demo (SMPLX)

Download Demo required data

1. Register on the BEDLAM2 website
2. Register on the CameraHMR website
3. Register on the SMPL-X website.
4. Run the following script:

bash scripts/fetch_demo_data_bedlam2.sh

Run Demo

Run the demo with the following command. It will run the demo on all images in the specified --image_folder, and save renderings of the reconstructions and the output mesh in --out_folder. Make sure to set --model_type to smplx.

python demo.py --image_folder demo_images --output_folder output_images_smplx --model_type smplx

Training and Evaluation (SMPLX)

Download Training required data

1. Register on the BEDLAM2 website.
2. Download and untar the BEDLAM2 30fps images (-png) in data/training-images/bedlam_v2 and the GT motion file b2_motions_npz_training.tar in data/training-labels/bedlam_v2 from the BEDLAM2 download page.
3. Run the following script to download extra necessary files. ```bash bash download_util_bedlam2.sh
4. Note that if you want to use BEDLAM1 data in training as well then you need to download the labels from BEDLAM website from the section SMPL-X ground truth labels compatible with BEDLAM2. After downloading them unzip in data/training-labels/bedlam-labels-v2-format

Training

Once the data is downloaded, you can run the training with the following command. We override the MODEL.TYPE to smplx to train an SMPLX model.

python train.py data=bedlam_v2_v1 experiment=bedlam2 exp_name=train_smplx_run1

Evaluation

Download Evaluation required data

To run the evaluation along with the checkpoints and SMPLX model files downloaded in Demo section you also need to download the test labels for 3DPW, EMDB, RICH from CameraHMR website using the following script. The following script also download some utilty files needed to run the evaluation.

bash scripts/fetch_test_labels.sh

Note: We cannot provide the original images for 3DPW, EMDB, RICH. These images must be obtained from their original sources.

The images could be downloaded and stored in data/test-images. This is the default directory structure for the images. If you have stored images at other location you could modify the path of the images here

├── 3DPW
│   └── imageFiles
├── EMDB
│   ├── P0
│   ├── P1
│   ├── P2
│   ├── P3
│   ├── P4
│   ├── P5
│   ├── P6
│   ├── P7
│   ├── P8
│   └── P9
├── RICH
│   └── test

Evaluation

For evaluation, use the following command.

python eval.py data=eval_smplx experiment=bedlam2