António Pinheiro: Universidade da Beira Interior, Portugal
Ricardo de Queiroz: Universidade Federal de Brasilia, Brasil
Stuart Perry: University of Technology Sydney, Australia
Luís Cruz: Instituto de Telecomunicações, Universidade de Coimbra, Portugal
Irene Viola: Centrum Wiskunde & Informatica, Netherlands
Submission Status: Open | Submission Deadline: 31 January 2024
EURASIP Journal on Image and Video Processing is calling for submissions to our Collection on Advances on Point Cloud technology: From coding and quality evaluation to applications
In this special issue of the EURASIP JIVP, leading researchers and practitioners in academia, industry and standard-bodies are invited to contribute to the advancement of the state of the art on point cloud processing technology, by submitting new coding models based either on the traditional approaches or in machine learning technology, new quality models for subjective quality estimation, new objective quality measures, and new methods and applications of point cloud processing.
Point Clouds (PC) are one of the formats being considered to represent plenoptic information. They provide a flexible representation of 3D visual information with multiple applications to virtual, mixed and augmented reality, computer graphics modelling, medical imaging, mobile robot environment modelling for use in autonomous driving and other applications.
However, the widespread adoption of point cloud applications faces several challenges, amongst which the most pressing one is the volume of data required to represent the geometry and attributes of the increasingly larger sets of points generated by point cloud capture and generation processes and devices.
Solving this problem calls for advanced coding models able to reduce the data size of point clouds to volumes compatible with efficient transmission, storage and processing.
In the last few years several point cloud coding methods have been proposed based on different representation principles, from octree decompositions to projection onto 2D images followed by waveform coding via legacy image coders. Recently a new class of methods leveraging the power of (deep) machine learning has been proposed, achieving good performance in geometry information coding, that is being followed by integrating machine learning based attribute compression tools. Besides the prospects of achieving coding performance (geometry and attributes) exceeding the state-of-art performance the machine learning solutions also permit direct compressed domain processing, with significant savings in the computational costs associated with decoding operations.
This Collection supports and amplifies research related to SDG 09.