Introduction
Summarize the problem, why it matters, what is missing in prior work, and the single-line insight behind your method. Keep it short and skimmable for readers outside your sub-field.
Key Contributions
- State your 2–4 crisp technical contributions. Avoid marketing; keep to verifiable claims.
- Highlight what is new compared to existing methods or baselines.
- Mention the settings and tasks where your method shows clear gains.
Method at a Glance
Inputs
Describe the signals you condition on (e.g., edges, masks, audio, text). Explain why they are sufficient and interpretable.
Backbone
Give a one-liner on the core architecture and how conditioning is injected (e.g., cross-attention, adapters, skip connections).
Objective
Explain the training target, losses, or supervision strategy, especially anything that enforces disentanglement or controllability.
Core Demo
Use this section to show why your chosen inputs/controls are informative and how the model recombines them to produce the output.
Replace the sample assets under public/images with your own inputs/intermediates/outputs.
Applications
Use this space to demonstrate how your representation plugs into downstream tasks (restoration, editing, generation, alignment). Swap the assets and copy to match your use cases.
A. Task A: Qualitative + Metrics
Show inputs, your output, and a few baselines. Add PSNR/SSIM/LPIPS or task-specific metrics below.
Baselines vs Ours
Poster
BibTeX
@inproceedings{Your2025Project,
title = {Project Name: Concise Claim in One Line},
author = {You, First and Collaborator, Second and Advisor, Senior},
booktitle = {Conference on Great Ideas},
year = {2025}
}