Summary of Point Cloud Compression with Bits-back Coding, by Nguyen Quang Hieu et al.
Point Cloud Compression with Bits-back Coding
by Nguyen Quang Hieu, Minh Nguyen, Dinh Thai Hoang, Diep N. Nguyen, Eryk Dutkiewicz
First submitted to arxiv on: 9 Oct 2024
Categories
- Main: Computer Vision and Pattern Recognition (cs.CV)
- Secondary: Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
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 This paper introduces a novel lossless compression method for compressing geometric attributes of point cloud data using bits-back coding and a deep learning-based probabilistic model. The method estimates the entropy of the point cloud information using a convolutional variational autoencoder (CVAE) and then compresses the geometric attributes with bits-back coding, capturing potential correlations between data points in a lower-dimensional latent space. The approach achieves a competitive compression ratio to conventional deep learning-based methods while reducing storage and communication costs. Comprehensive evaluations show that the cost of the overhead is significantly small compared to the reduction in compression ratio when compressing large point cloud datasets. The method can achieve an average compression ratio of 1.56 bit-per-point, outperforming Google’s Draco with a compression ratio of 1.83 bit-per-point. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper finds a new way to make big files smaller using special math and computer learning techniques. It works by first understanding how the data is organized, then using that information to shrink it down without losing any important details. The result is a much more efficient way of storing and sending large files, which could be very useful in many different areas. |
Keywords
» Artificial intelligence » Deep learning » Latent space » Probabilistic model » Variational autoencoder
