Best Part — Our model works with quick doodles! We are not expecting you to spend hours sketching.
Most methods claim to work with sketches, but they're actually trained on edge maps; perfect, mechanical outlines extracted from photos. Real human sketches are messy, abstract, and beautiful.
📷 Reference Photo
The original image
❌ Edge Map (Canny)
Pixel-perfect boundaries extracted automatically
✓ Freehand Sketch
Abstract, distorted, but full of semantic meaning
Our breakthrough: We train AI to understand the semantic intent behind your rough strokes, not pixel-perfect alignment.
Hover to Transform
Watch rough sketches become photorealistic images. Hover on any card to see the AI-generated results cycle through!
Sketch
"A bench in the garden"
Sketch
"Train going on the track"
Sketch
"A man is flying kite in the sky"
Sketch
"A girl is sitting on a horse"
Sketch
"Airplane is standing on the airport"
Sketch
"Three people walking with umbrellas"
Sketch
"Two people at a food table outdoors"
Sketch
"Two girls playing with frisbee"
Sketch
"A building with a clock on it"
Sketch
"Two giraffes standing in the zoo"
State-of-the-Art Results
Evaluated on 475 freehand sketches from the FS-COCO dataset
How We Compare to Others
See the difference between our method and state-of-the-art baselines on the same freehand sketches.
"A girl is sitting on a horse"
"A bench in the garden"
Our Approach
A modulation-based method that prioritizes semantic understanding over pixel alignment.
Semantic Sketch Features
We leverage a CLIP-based encoder fine-tuned for freehand sketches to capture high-level semantic information.
Modulation Network
Instead of direct conditioning, we modulate the diffusion process using learned scale and shift maps.
Attention Supervision
A novel loss function that enables training without pixel-aligned ground truth images.
BibTeX
@misc{bourouis2026sketchingrealityfreehandscenesketches,
title={SketchingReality: From Freehand Scene Sketches To Photorealistic Images},
author={Ahmed Bourouis and Mikhail Bessmeltsev and Yulia Gryaditskaya},
year={2026},
eprint={2602.14648},
archivePrefix={arXiv},
primaryClass={cs.CV},
url={https://arxiv.org/abs/2602.14648}
}