Ampli

The goal of AMPLI is to propose high-level techniques that assist artists in authoring and creating virtual worlds. The project develops tools that support users in their creative tasks while preserving their artistic freedom and their control at several levels of detail.

It relies on machine learning methods able to generate detailed maps from sketches or from coarse input data, so that a rough intention can be amplified into a complete, coherent scene.

Publications

2026 EcoViz: a tool for visual analysis and photorealistic rendering of forest landscape model simulations — Ecography
2025 Vector-based Terrain Modelling — Computer Graphics Forum
2024 Deadwood — ACM Transactions on Graphics
2024 Terrain Amplification using Multi-scale Erosion — ACM Transactions on Graphics
2024 Vector-based terrain authoring — Journées françaises d'informatique graphique (jFIG) 2024
2023 Authoring Terrains with Spatialised Style — Computer Graphics Forum
2023 Forming Terrains by Glacial Erosion — ACM Transactions on Graphics
2023 Interactive Authoring of Terrain using Diffusion Models — Computer Graphics Forum
2022 Gradient Terrain Authoring — Computer Graphics Forum
2022 PTRM: Perceived Terrain Realism Metric — ACM Transactions on Applied Perception
2022 StyleDEM: a Versatile Model for Authoring Terrains — Journées françaises d'informatique graphique (jFIG) 2022 à Bordeaux
2021 Synthesizing Geologically Coherent Cave Networks — Computer Graphics Forum
2021 Synthesizing Geologically Coherent Cave Networks — Journées Françaises de l'Informatique Graphique JFIG 2021
2020 Data-driven Authoring of Large-scale Ecosystems — ACM Transactions on Graphics
2020 Modeling Rocky Scenery using Implicit Blocks — Journées Françaises d'Informatique Graphique (JFIG 2020)