Jiahao Zhu, Zixuan Chen, Guangcong Wang, Xiaohua Xie✉️, Yi Zhou✉️

Recent advancements in text-to-3D generation improve the visual quality of Score Distillation Sampling (SDS) and its variants by directly connecting Consistency Distillation (CD) to score distillation. However, due to the imbalance between self-consistency and cross-consistency, these CD-based methods inherently suffer from improper conditional guidance, leading to sub-optimal generation results. To address this issue, we present SegmentDreamer, a novel framework designed to fully unleash the potential of consistency models for high-fidelity text-to-3D generation. Specifically, we reformulate SDS through the proposed Segmented Consistency Trajectory Distillation (SCTD), effectively mitigating the imbalance issues by explicitly defining the relationship between self- and cross-consistency. Moreover, SCTD partitions the Probability Flow Ordinary Differential Equation (PF-ODE) trajectory into multiple sub-trajectories and ensures consistency within each segment, which can theoretically provide a significantly tighter upper bound on distillation error. Additionally, we propose a distillation pipeline for a more swift and stable generation. Extensive experiments demonstrate that our SegmentDreamer outperforms state-of-the-art methods in visual quality, enabling high-fidelity 3D asset creation through 3D Gaussian Splatting (3DGS).

conda create -n sctd python=3.9.16 cudatoolkit=11.8
conda activate sctd
pip install -r requirements.txt
pip install submodules/diff-gaussian-rasterization/
pip install submodules/simple-knn/
pip install submodules/point-e/
pip install tensorboard

# train script
python train.py --opt ./config/bagel.yaml --guidance_type sctd --phase_num 5 --num_ddim_timesteps 50 --guid_scale 7.5

@InProceedings{Zhu_2025_ICCV,
    author    = {Zhu, Jiahao and Chen, Zixuan and Wang, Guangcong and Xie, Xiaohua and Zhou, Yi},
    title     = {SegmentDreamer: Towards High-fidelity Text-to-3D Synthesis with Segmented Consistency Trajectory Distillation},
    booktitle = {Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV)},
    month     = {October},
    year      = {2025},
    pages     = {15864-15874}
}

Our work is built upon LucidDreamer and Connecting Consistency Distillation to Score Distillation for Text-to-3D Generation. Thanks for their contributions to the 3D community!