Genesis

A Generative Engine for Hierarchical Satellite Image Synthesis — ACM SIGSPATIAL 2026 (Oral).

Genesis completes a sparse set of seed satellite tiles into a full, zoomable Web-Mercator quadtree — filling every scale and location. Two flow-matching JiT operators do the work: a parent-conditional super-resolution model (one 256×256 parent tile → the 512×512 mosaic of its four children, one zoom level down) and a mask-based outpainting model (256×256 completion under quadrant masks). Composed into a pyramid engine, they bring the seeds to a common working level, greedily outpaint to complete that level, super-resolve down to the leaf tiles, and fill the coarser levels by downsampling — yielding a pyramid that is consistent across scales and honors every seed. Models are trained on the Git-10M global tile corpus.

Subash Khanal, Yangzhi Cui, Daniel Cher, Eric Xing, Brian Wei, Srikumar Sastry, Nathan Jacobs — Washington University in St. Louis.

Checkpoints

All models sample with 50 flow-matching steps, cfg = 1.0 (the paper eval setting) and load EMA weights. Main runs are trained to 800k steps. The main SR models are DINOv3-conditioned (+LPIPS); the main outpainting models are no-DINO.

Path	Model / arch	Size	Train steps	What it's for
main/superresolution/main_B16/sr-full-tile-stage3-step0800000.ckpt	SR · JiT-B/16	3.3 GB	800k	Main SR (DINOv3 + LPIPS), 256→512
main/superresolution/main_H16/sr-full-tile-stage3-step0800000.ckpt	SR · JiT-H/16	23 GB	800k	Main SR, largest model
main/outpainting/main_B16/op-new-stage3-step0800000.ckpt	OP · JiT-B/16	3.6 GB	800k	Main outpainting (no-DINO, quadrant masks)
main/outpainting/main_H16/op-new-stage3-step0800000.ckpt	OP · JiT-H/16	26 GB	800k	Main outpainting, largest model
shared/git10m_quad_meta.json	metadata	1.7 GB	—	Git-10M split metadata (paper eval)
shared/hierarchy_index_quad.pkl	metadata	0.6 GB	—	Parent↔child quadtree index (paper eval)
dinov3_vitl16_pretrain_sat493m-eadcf0ff.pth	DINOv3 ViT-L/16 (SAT-493M)	1.2 GB	—	Frozen conditioning encoder for SR
example_sr_op.py	script	—	—	Runnable quick start (below)

The ablations/ folder holds additional checkpoints used only to reproduce the paper's ablation table (see the code repo's docs/EVALUATION.md).

Quick start

Install the code repo (checkpoints download from this HF repo automatically):

git clone https://github.com/mvrl/genesis && cd genesis
uv sync && source .venv/bin/activate    # Python 3.13, torch 2.8 cu128

Then run one SR ×2 pass and one quadrant-outpainting pass on a live Esri World Imagery tile (the Sydney Opera House waterfront at zoom 16 serves as the dummy input — the script fetches it over the network):

import os, sys
sys.path.insert(0, "src")
sys.path.insert(0, "demos")   # demos/ last = highest priority (its utils/ package must win)

from huggingface_hub import hf_hub_download
from PIL import Image
from utils.genesis_common import (
    fetch_tile_at_zoom, inference_device,
    load_sr_denoiser, load_op_denoiser,
    run_superresolution, run_outpainting,
)

device = inference_device()   # cuda if available, else cpu

sr_ckpt = hf_hub_download("MVRL/genesis", "main/superresolution/main_B16/sr-full-tile-stage3-step0800000.ckpt")
op_ckpt = hf_hub_download("MVRL/genesis", "main/outpainting/main_B16/op-new-stage3-step0800000.ckpt")
dino    = hf_hub_download("MVRL/genesis", "dinov3_vitl16_pretrain_sat493m-eadcf0ff.pth")

lat, lon, zoom = -33.8568, 151.2153, 16
tile = fetch_tile_at_zoom(lon, lat, zoom)                 # one real 256×256 Esri tile

# SR: z16 parent → 512×512 mosaic of its four z17 children (steps=50, cfg=1.0 defaults)
sr_model, _ = load_sr_denoiser(sr_ckpt, "JiT-B/16", dino_weights=dino, device=device)
run_superresolution(sr_model, tile, device, target_zoom=zoom + 1).save("genesis_sr_out.png")

# OP: keep the upper-left quadrant real, outpaint the other three (white = hole)
mask = Image.new("L", (256, 256), 255)
mask.paste(0, (0, 0, 128, 128))
op_model, _ = load_op_denoiser(op_ckpt, "JiT-B/16", dino_weights="", device=device)
run_outpainting(op_model, tile, mask, device, tile_zoom=zoom).save("genesis_op_out.png")

Or as a one-liner — download example_sr_op.py from this repo into the checkout root and run:

python example_sr_op.py

Live imagery: Esri World Imagery — Source: Esri, Maxar, Earthstar Geographics, and the GIS User Community.

Demos

Interactive Gradio apps live in the code repo — see demos/README.md. The main one is the full-pyramid builder (demos/app_pyramid_unified.py): click a satellite map to drop sparse seed tiles and watch Genesis complete an entire 4-level quadtree live, tile-by-tile, with a stitched-PNG download. Companion apps cover single-tile 256→512 SR, quadrant outpainting, a large lateral-outpainting grid, a deep 6-level pyramid, and no-model interpolation baselines.

Reproduce the paper

One command per paper table — see docs/EVALUATION.md in the code repo: bash src/eval/run_all_sr.sh · run_all_op.sh · run_all_ablation.sh · run_all_dense500.sh, each with the paper-exact defaults baked in.

Citation

@inproceedings{khanal2026genesis,
  title     = {{Genesis}: A Generative Engine for Hierarchical Satellite Image Synthesis},
  author    = {Khanal, Subash and Cui, Yangzhi and Cher, Daniel and Xing, Eric and
               Wei, Brian and Sastry, Srikumar and Jacobs, Nathan},
  booktitle = {ACM SIGSPATIAL International Conference on Advances in
               Geographic Information Systems (SIGSPATIAL)},
  year      = {2026},
}

License & acknowledgments

The Genesis code and checkpoints are released under the MIT License (see the code repo). Both Genesis operators build on JiT — we thank the authors for releasing their codebase.

dinov3_vitl16_pretrain_sat493m-eadcf0ff.pth is the satellite-pretrained (SAT-493M) ViT-L/16 encoder from Meta AI's DINOv3 release, redistributed here unmodified for convenience; it remains governed by Meta's DINOv3 license, not MIT. At inference the SR model loads it via torch.hub.load('facebookresearch/dinov3', 'dinov3_vitl16', weights=...).

Training data comes from the Git-10M global satellite-tile corpus. Demo/example imagery is fetched live from Esri World Imagery (attribution above).