ModelPublicadoExterno

DermDepth

This repository contains model checkpoints from the paper DermDepth: Toward Monocular Metric Scale 3D Reconstruction Models for Dermatology (Carrión & Norouzi, MICCAI 2026).

Fuentehuggingface.co/hcarrion/DermDepth/tree/e384b43821a14f69a1feb6258ea7ab43f9ad1d8d

TipoOtherModelo baseRuicheng/moge-2-vitl-normalVersiónDerm Depth Synth NormalsCcalidad

Fuente del modelo

Tipo: Other
Tamaño del archivo: 1,23 GB
Modelo base: Ruicheng/moge-2-vitl-normal
Versión: vDerm Depth Synth Normals
Actualizado: 10 oct 2026
Fuente: Hugging Face

Descripción de la fuente

DermDepth: Monocular Metric Scale 3D Reconstruction for Dermatology

This repository contains model checkpoints from the paper DermDepth: Toward Monocular Metric Scale 3D Reconstruction Models for Dermatology (Carrión & Norouzi, MICCAI 2026).

Checkpoints

Filename	Training data	Notes
`DermDepth_Synth.pt`	D-Synth only	"DermDepth_S" in the paper. Synthetic-only baseline.
`DermDepth_Synth_SKINL2_WoundsDB.pt`	D-Synth → SKINL2 + WoundsDB	Intermediate stage (before DDI pseudo-GT).
`DermDepth_Synth_SKINL2_WoundsDB_DDI.pt`	D-Synth → SKINL2 + WoundsDB → DDI pseudo-GT	Best model. Corresponds to "DermDepth" in the paper.
`DermDepth_Synth_Normals.pt`	D-Synth, normal-head emphasis	Trained normal-head model.

Key results (held-out test sets, paper Table 1)

Method	SKINL2 Scale	WoundsDB Scale	DDI Ratio	Fitzpatrick Disparity
MoGe-2 (baseline)	16.10×	0.62×	81.0×	10.90
DermDepth_Synth	1.11×	0.28×	9.2×	1.70
DermDepth (best)	0.87×	0.91×	1.95×	1.02

Scale ratio target is 1.0×. See the paper for full benchmarks and SI-δ1 details.

Usage

These checkpoints are designed to be loaded by MoGe-2, modified per the DermDepth code repository. The repo contains end-to-end training, inference, and evaluation scripts.

Quick download:

from huggingface_hub import hf_hub_download
ckpt = hf_hub_download(
    repo_id="hcarrion/DermDepth",
    filename="DermDepth_Synth_SKINL2_WoundsDB_DDI.pt",
)

Citation

@inproceedings{carrion2026dermdepth,
  title     = {DermDepth: Toward Monocular Metric Scale 3D Reconstruction Models for Dermatology},
  author    = {Carri{\'o}n, H{\'e}ctor and Norouzi, Narges},
  booktitle = {Medical Image Computing and Computer-Assisted Intervention (MICCAI)},
  year      = {2026}
}

License

CC BY-NC 4.0 (research / non-commercial use). The base MoGe-2 weights remain under their original license — see the MoGe repository for details.

Related Resources

Code & training scripts: https://github.com/hectorcarrion/dermdepth
D-Synth dataset (training data): https://huggingface.co/datasets/hcarrion/D-Synth
Base model: https://huggingface.co/Ruicheng/moge-2-vitl-normal

Fuentes

1 fuente

Hugging FacePrincipalVersión e384b43821a14f69a1feb6258ea7ab43f9ad1d8d

huggingface.co/hcarrion/DermDepth/tree/e384b43821a14f69a1feb6258ea7ab43f9ad1d8d

Verificado 10 oct

Download files

Artefactos del modelo

4 artefactos

DermDepth_Synth_Normals.pt

pt · 1,23 GB · SHA-256 cc21e80ebf43…ebe0 · Hugging Face

Download file

DermDepth_Synth_SKINL2_WoundsDB_DDI.pt

pt · 1,23 GB · SHA-256 cce1c836802f…b5ba · Hugging Face

Download file

DermDepth_Synth_SKINL2_WoundsDB.pt

pt · 1,23 GB · SHA-256 89e6aa75b481…828e · Hugging Face

Download file

DermDepth_Synth.pt

pt · 1,23 GB · SHA-256 6cf834c88469…1390 · Hugging Face

Download file

Extractos de fuentes

2 extractos

Readme1 agohuggingface.co/hcarrion/DermDepth

--- base_model: Ruicheng/moge-2-vitl-normal library_name: pytorch license: cc-by-nc-4.0 pipeline_tag: depth-estimation tags: - depth-estimation - dermatology - 3d-reconstruction - medical-imaging - skin-lesion - moge --- # DermDepth: Monocular Metric Scale 3D Reconstruction for Dermatology This repository contains model checkpoints from the paper [DermDepth: Toward Monocular Metric Scale 3D Reconstruction Models for Dermatology](https://huggingface.co/papers/2607.13010) (Carrión & Norouzi, MICCAI 2026). ## Checkpoints | Filename | Training data | Notes | |---|---|---| | `DermDepth_Synth.pt` | D-Synth only | "DermDepth_S" in the paper. Synthetic-only baseline. | | `DermDepth_Synth_SKINL2_WoundsDB.pt` | D-Synth → SKINL2 + WoundsDB | Intermediate stage (before DDI pseudo-GT). | | `DermDepth_Synth_SKINL2_WoundsDB_DDI.pt` | D-Synth → SKINL2 + WoundsDB → DDI pseudo-GT | **Best model. Corresponds to "DermDepth" in the paper.** | | `DermDepth_Synth_Normals.pt` | D-Synth, normal-head emphasis | Trained normal-head model. | ## Key results (held-out test sets, paper Table 1) | Method | SKINL2 Scale | WoundsDB Scale | DDI Ratio | Fitzpatrick Disparity | |---|:---:|:---:|:---:|:---:| | MoGe-2 (baseline) | 16.10× | 0.62× | 81.0× | 10.90 | | DermDepth_Synth | 1.11× | 0.28× | 9.2× | 1.70 | | **DermDepth (best)** | **0.87×** | **0.91×** | **1.95×** | **1.02** | Scale ratio target is 1.0×. See the paper for full benchmarks and SI-δ₁ details. ## Usage These checkpoints are designed to be loaded by [MoGe-2](https://github.com/microsoft/MoGe), modified per the [DermDepth code repository](https://github.com/hectorcarrion/dermdepth). The repo contains end-to-end training, inference, and evaluation scripts. Quick download: ```python from huggingface_hub import hf_hub_download ckpt = hf_hub_download( repo_id="hcarrion/DermDepth", filename="DermDepth_Synth_SKINL2_WoundsDB_DDI.pt", ) ``` ## Citation ```bibtex @inproceedings{carrion2026dermdepth, title = {DermDepth: Toward Monocular Metric Scale 3D Reconstruction Models for Dermatology}, author = {Carri{\'o}n, H{\'e}ctor and Norouzi, Narges}, booktitle = {Medical Image Computing and Computer-Assisted Intervention (MICCAI)}, year = {2026} } ``` ## License CC BY-NC 4.0 (research / non-commercial use). The base MoGe-2 weights remain under their original license — see [the MoGe repository](https://github.com/microso...

Note1 agohuggingface.co/hcarrion/DermDepth

Source context: 0 downloads · 0 likes · Pipeline depth-estimation · Library pytorch · Repo hcarrion/DermDepth