Release Code for 'Clip Fusion with Bi-level Optimization for Human MeshReconstruction from Monocular Videos' that Was Accepted by ACM MM(2023).

Data Download

• Please refer to this repo to download the HMR datasets.

Training

• You can use the below command to train our code.

  python train.py --cfg ./config/{config_name}.yaml

---

Some Demo Videos For 'Clip Fusion with Bi-level Optimization for Human MeshReconstruction from Monocular Videos'

New Demo on Internet(top->ours, mid->mpsnet, bottom->tcmr)

demo_new.mp4

Demo 1(internet) on TCMR

tcmr_demo1.mp4

Demo 1(internet) on BiCF(ours)

bicf_demo1.mp4

Demo 2(internet) on TCMR

Demo 2(internet) on BiCF(ours)

Demo 3(internet) on TCMR

Demo 3(internet) on BiCF(ours)

Demo 4(internet) on TCMR

Demo 4(internet) on BiCF(ours)

Demo 5(3dpw) on TCMR

tcmr_demo5.mp4

Demo 5(3dpw) on BiCF(ours)

bicf_demo5.mp4

---

Licence

• This project is licensed under the terms of the MIT license.

---

Related Works

TCMR VIBE