Had the pleasure of (virtually) hosting @kvuongdev today at my lab and hear about his awesome work! Thanks for sharing your insights from working on 3D and video models.
🎉 Excited to introduce TRON, a relighting framework for 3D captures.
💡TRON pairs a neural renderer with 3D Gaussian reconstructions, achieving realistic quality, with 3D, material, & lighting control at interactive frame rates.
arxiv.org/abs/2606.11314research.nvidia.com/labs/sil/proje…
SpectralSplats is accepted at #ECCV2026! 🎉 Tracking 3DGS across frames is harder than it looks — appearance losses break the moment the pose drifts. Spectral moments keep it robust.
@eccvconf
1/9 Excited to share SpectralSplats! 📢 Given a 3DGS asset + target video, we deform it to match the video via differentiable rendering. Appearance-based tracking fails when the initial pose is even slightly off. Our spectral loss stays robust.
🔗 avigailco.github.io/SpectralSplats
🧵
Excited that RadarGen is accepted to #ECCV2026! Looking forward to chatting radar simulation in Malmö — let’s push this underexplored topic forward.
@eccvconf
🚗📡Radar is the unsung hero of AV perception: widespread in cars, yet overlooked in simulation.
Introducing RadarGen: Realistic radar synthesis from cameras using diffusion.
Massive kudos to my fantastic team at @TechnionLive and @NVIDIAAIradargen.github.io
📢 New paper: FlowBender.
Conditional generators drift from their own conditioning. The usual fix: tune guidance & pray 🙏
💡We train them to self-correct from their own error
+4.5 dB on 3D texturing, +4.9 dB on SR.
Give it a spin and bend away👇
flow-bender.github.io
Conditional diffusion/flow models often produce outputs inconsistent with the very signal conditioning them. The error is easily measurable, yet models are never trained to act on it.
In FlowBender (now on arXiv), we train the model to correct its own errors. 🧵
