Hi3D: Pursuing High-Resolution Image-to-3D Generation with Video Diffusion Models

This is the official implementation of our paper Hi3D: Pursuing High-Resolution Image-to-3D Generation with Video Diffusion Models by Haibo Yang, Yang Chen, Yingwei Pan, Ting Yao, Zhineng Chen, Chong-Wah Ngo, Tao Mei .

Abstract

Despite having tremendous progress in image-to-3D generation, existing methods still struggle to produce multi-view consistent images with high-resolution textures in detail, especially in the paradigm of 2D diffusion that lacks 3D awareness. In this work, we present High-resolution Image-to-3D model (Hi3D), a new video diffusion based paradigm that redefines a single image to multi-view images as 3D-aware sequential image generation (i.e., orbital video generation). This methodology delves into the underlying temporal consistency knowledge in video diffusion model that generalizes well to geometry consistency across multiple views in 3D generation. Technically, Hi3D first empowers the pre-trained video diffusion model with 3D-aware prior (camera pose condition), yielding multi-view images with low-resolution texture details. A 3D-aware video-to-video refiner is learnt to further scale up the multi-view images with high-resolution texture details. Such high-resolution multi-view images are further augmented with novel views through 3D Gaussian Splatting, which are finally leveraged to obtain high-fidelity meshes via 3D reconstruction. Extensive experiments on both novel view synthesis and single view reconstruction demonstrate that our Hi3D manages to produce superior multi-view consistency images with highly-detailed textures.

Demo

🌟 Hi3D-codes

🎉🎉🎉 We have released the training code for the first and second stages. You can easily modify our code to finetune Stabel Video Diffusion for the image-to-video task (first stage) and the video-to-video task (second stage).

Official codes for ACM MM24 paper "Hi3D: Pursuing High-Resolution Image-to-3D Generation with Video Diffusion Models"

• First stage checkpoint release. The checkpoint is available at here.
• First stage inference codes.
• Second stage checkpoint release. The checkpoint is available at here.
• Second stage inference codes.
• Training codes and datasets.

Setup

1. git clone this repo

cd ~
git clone https://github.com/yanghb22-fdu/Hi3D-Official.git

2. Install latest conda or miniconda

#Example On Linux
mkdir -p ~/miniconda3
cd ~/miniconda3	
wget https://repo.anaconda.com/miniconda/Miniconda3-latest-Linux-x86_64.sh -O ~/miniconda3/miniconda.sh
bash ~/miniconda3/miniconda.sh -b -u -p ~/miniconda3
rm -rf ~/miniconda3/miniconda.sh

3. Install conda requirements

cd ~/Hi3D-Official
~/miniconda3/bin/conda env create --file environments.yaml --prefix .conda -y -q

4. Download weights/models/decoder

cd ~/Hi3D-Official
curl https://huggingface.co/hbyang/Hi3D/resolve/main/ckpts.zip -L -o ckpts.zip
unzip ckpts.zip
cd ckpts
curl https://huggingface.co/stabilityai/stable-video-diffusion-img2vid-xt/resolve/main/svd_xt_image_decoder.safetensors -L -o svd_xt_image_decoder.safetensors

Inference

1. Download stage checkpoints

cd ~/Hi3D-Official/ckpts
curl https://huggingface.co/hbyang/Hi3D/resolve/main/first_stage.pt -L -o first_stage.pt
curl https://huggingface.co/hbyang/Hi3D/resolve/main/second_stage.pt -L -o second_stage.pt

2. Make sure you have the following models.

Hi3D-Official
|-- ckpts
    |-- metric_models
    |-- dpt_hybrid_384.pt
    |-- first_stage.pt
    |-- ViT-L-14.ckpt
    |-- second_stage.pt
    |-- open_clip_pytorch_model.bin

3. Run Hi3D to produce multiview-consistent images.

First stage

CUDA_VISIBLE_DEVICES=0 python pipeline_i2v_eval_v01.py \
    --denoise_checkpoint "ckpts/first_stage.pt" \
    --image_path "demo/3.png" \
    --output_dir "outputs/3"

Second stage

CUDA_VISIBLE_DEVICES=0 python pipeline_i2v_eval_v02.py \
    --denoise_checkpoint "ckpts/second_stage.pt" \
    --image_path "demo/3.png" \
    --output_dir "outputs/3"

Training

1. We only provide an example dataset here. Download the example dataset and unzip. To create your own dataset from 3d models, you can use our modified version of Syncdreamer's blender render utility script. This script supports the usdz/fbx/glb file format. See the script for instructions on how to generate the dataset.

2. First stage training: Modify the model.ckpt_path in the train-v01.yaml file to point to the location where you downloaded the checkpoint.

python train_ddp_spawn.py \
    --base configs/train-v01.yaml \
    --no-test True \
    --train True \
    --logdir outputs/logs/train-v01 \
    --outckpt ckpts

3. Prepare stage 2. Modify 'ckpt_path' in train-v02.yaml to point to the ckpt generated with first stage. tool_make_init_svd_to_vid2vid.py will output the new safetensor in the ckpts folder and generate a modified train-v02.yaml with its path.

### modify svd to fit our config
python tool_make_init_svd_to_vid2vid.py

3. Second stage training: modify train-v02-edited.yaml to fit your need.

### training
python train_ddp_spawn.py \
    --base configs/train-v02-edited.yaml \
    --no-test True \
    --train True \
    --logdir outputs/logs/train-v02 \
    --outckpt ckpts

Acknowledgement

The Hi3D-Diffusion code is heavily based on the generative-models project.

Citation

@inproceedings{yang2024hi3d,
  title={Hi3D: Pursuing High-Resolution Image-to-3D Generation with Video Diffusion Models},
  author={Haibo Yang and Yang Chen and Yingwei Pan and Ting Yao and Zhineng Chen and Chong-Wah Ngo and Tao Mei},
  booktitle={ACM MM},
  year={2024}
}