ImportantArticles

ID	Year	Name	Note	Tags	Link
	2023	NeuralLift-360: Lifting An In-the-wild 2D Photo to A 3D Object with 360° Views	SDS + Fine-tuned CLIP text embedding + Depth supervision> ✅ 整体上是类似 SDS 的优化方法，再结合其它的损失函数。 ✅ (1) 渲染不同视角，并对渲染结果用 clip score打分。 ✅ (2) 监督深度信息。
	2023	Zero-1-to-3: Zero-shot One Image to 3D Object	Generate novel view from 1 view and pose, with 2d model. Then, run SJC / SDS-like optimizations with view-conditioned model. ✅ (1) 用 2D diffusion 生成多视角。用 SDS 对多视角图像生成3D．
	2024	CAT3D

本文出自CaterpillarStudyGroup，转载请注明出处。

https://caterpillarstudygroup.github.io/ImportantArticles/