Better 3D reconstruction without exact model match by using database of patches

Better 3D reconstruction without exact model match by using database of patches

Patch2CAD: Patchwise Embedding Learning for In-the-Wild Shape Retrieval from a Single Image

arXiv paper abstract https://arxiv.org/abs/2108.09368

arXiv PDF paper https://arxiv.org/pdf/2108.09368.pdf

3D perception of object shapes from RGB image input is fundamental towards semantic scene understanding

... To achieve a mapping between image views of objects and 3D shapes, we leverage CAD model priors from existing large-scale databases, and propose a novel approach towards constructing a joint embedding space between 2D images and 3D CAD models in a patch-wise fashion

-- establishing correspondences between patches of an image view of an object and patches of CAD geometry.

This enables part similarity reasoning for retrieving similar CADs to a new image view without exact matches in the database.

... approach is more robust than state of the art in real-world scenarios without any exact CAD matches.

Please like and share this post if you enjoyed it using the buttons at the bottom! Stay up to date. Subscribe to my posts https://morrislee1234.wixsite.com/website/contact Web site with my other posts by category https://morrislee1234.wixsite.com/website #ComputerVision #3D #AINewsClips #AI #ML #ArtificialIntelligence #MachineLearning