Researcher flags Depth Pro 2 border artifacts: matting vs depth confusion, RL could fix it
giffmana · x · 2026-11-21
Lucas Beyer (giffmana) critiqued Depth Pro 2's depth visualizations: the wide green/yellow gradient at object borders would become a "wall" in 3D, suggesting the authors confused what's good for matting with what's good for depth. Correct output should jump straight from red to blue with no gradient. He notes DepthAnythingV2 is better here but also flawed, and argues this is a perfect RL problem — he wanted to tackle it for his RL-for-vision paper but ran out of time, inviting others to solve it and cite the paper.
More from Models
- Mystery stealth model appears; Reddit users guess it's 'Fable 5.5' — cryptomaniac1729 · 2026-11-21
- GPT-6.2 Astra tops AA-Robotics v0.1 with 75% fewer arm commands than Claude — ArtificialAnlys · 2026-11-21
- ChatGPT's PC use only works in new chat sessions, fails silently in old ones — LivingHighAndWise · 2026-11-21
- Qwen 27B quant runs on 8GB VRAM: 200k context, 92.5% Top-1 vs BF16 — Brief-Tap-6616 · 2026-11-21
- Rumor: DeepMind rebuilds Gemini architecture every 6 months with 3 parallel teams, ~6 releases a year — DeryaTR_ · 2026-11-21
- Google launches Gemini 4 Hydrogen: a speed-focused everyday model now live in Gemini app and Search — GeminiApp · 2026-11-21