Summary of Pickscan: Object Discovery and Reconstruction From Handheld Interactions, by Vincent Van Der Brugge et al.
PickScan: Object discovery and reconstruction from handheld interactions
by Vincent van der Brugge, Marc Pollefeys, Joshua B. Tenenbaum, Ayush Tewari, Krishna Murthy Jatavallabhula
First submitted to arxiv on: 17 Nov 2024
Categories
- Main: Computer Vision and Pattern Recognition (cs.CV)
- Secondary: Artificial Intelligence (cs.AI); Graphics (cs.GR); Machine Learning (cs.LG); Robotics (cs.RO)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary The proposed method enables the reconstruction of compositional 3D representations of scenes by allowing users to move around objects with an RGB-D camera, manipulate object positions, and output unique 3D models for each held-up object. This approach addresses limitations in existing methods that rely on strong appearance priors or do not support object manipulation. The method detects user-object interactions and extracts object masks, demonstrating a precision of 78.3% at 100% recall on a custom-captured dataset. Compared to the Co-Fusion baseline, this approach reduces chamfer distance by 73% while detecting significantly fewer false positives. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary The paper helps robots and augmented reality devices understand scenes better. It allows users to hold up objects with an RGB-D camera and creates a 3D model for each object. This is important because many current methods only work well on specific types of objects or don’t let you manipulate the objects. The new method is good at detecting which objects are being held up and reconstructing their shapes. It performs better than other similar methods in detecting objects correctly while reducing mistakes. |
Keywords
» Artificial intelligence » Precision » Recall
