Abstract:Convex polytopes have compact representations and exhibit convexity, which makes them suitable for abstracting obstacle-free spaces from various environments. Existing methods for generating convex polytopes always struggle to strike a balance between two requirements, producing high-quality polytope and efficiency. Moreover, another crucial requirement for convex polytopes to accurately contain certain seed point sets, such as a robot or a front-end path, is proposed in various tasks, which we refer to as manageability. In this paper, we show that we can achieve generation of high-quality convex polytope while ensuring both efficiency and manageability simultaneously, by introducing Fast Iterative Regional Inflation (FIRI).FIRI consists of two iteratively executed submodules: Restrictive Inflation (RsI) and computation of the Maximum Volume Inscribed Ellipsoid (MVIE) of convex polytope. By explicitly incorporating constraints that include the seed point set, RsI guarantees manageability. Meanwhile, the iterative monotonic optimization of MVIE, which serves as a lower bound of the volume of convex polytope, ensures high-quality results of FIRI. In terms of efficiency, we design methods tailored to the low-dimensional and multi-constrained nature of both modules, resulting in orders of magnitude improvement compared to generic solvers. Notably, for 2-D MVIE, we present a novel analytical algorithm that achieves linear-time complexity for the first time, further enhancing the efficiency of FIRI in the 2-D scenario. Extensive benchmarks conducted against state-of-the-art methods validate the superior performance of FIRI in terms of quality, manageability, and efficiency. Furthermore, various real-world applications showcase the generality and practicality of FIRI. The high-performance code of FIRI will be open-sourced for the reference of the community.

Efficient 3D region guarding for multimedia data processing

Efficient screen space approach for high complex dynamic scenes in real-time rendering

Efficient Density Control for 3D Gaussian Splatting

Hierarchical Video Data Modeling and Indexing for Virtual Scene Construction

A Locatable Zero Watermarking Scheme and Visualization for 3D Mesh Models

A novel segmentation algorithm for complex 3D mesh model in computer vision.

THREE-DIMENSIONAL ANIMATION COMPRESSION BASED ON SPACE-TIME SEGMENTATION

High Capacity Reversible Data Hiding for 3D Meshes in the PVQ Domain

Video abstraction based on the visual attention model and online clustering

Spatio-Temporal Video Segmentation of Static Scenes and Its Applications

Fast Automatic Elimination of Vertical Parallax of Multiview Images

Balancing Cpu And Gpu: Real-Time Visualization Of Large Scale 3d Scanning Models

Efficient View-Dependent LOD Control for Large 3D Unclosed Mesh Models of Environments

SEG-MAT: 3D Shape Segmentation Using Medial Axis Transform

Modeling Background and Segmenting Moving Objects from Compressed Video

Background Modeling Based On Region Segmentation

Optimally Guarding Perimeters and Regions with Mobile Range Sensors

Secure 3D medical Imaging

Fast Iterative Region Inflation for Computing Large 2-D/3-D Convex Regions of Obstacle-Free Space

Efficient Region of Visual Interests Search for Geo-multimedia Data

Efficient Streaming of Point-Based Models for Interactive 3-D Geometry Transmission