- PAMI2024
- mmWave
- conference
- mmWave
•
•
•
-
Scharstein-Szeliski Disparity Benchmark
3D surface modeling is a crucial task in various fields, including engineering, manufacturing, and medical imaging. Advancements in image, infrared (IR), and laser technologies have made it possible to develop more efficient and accurate 3D surface modeling techniques. Image-based surface modeling involves capturing images of surfaces from different perspectives and then reconstructing the 3D geometry from those images. IR-based surface modeling uses IR sensing technology to detect surface features such as textures, patterns, and colors. Laser-based surface modeling involves using lasers to measure the properties of surfaces with high precision and accuracy. The Scharstein-Szeliski Disparity Benchmark (SSDB) is a widely used dataset for evaluating the performance of 3D surface reconstruction algorithms.
-
Robust Multi-View Stereopsis
Robust Multi-View Stereopsis (RMS) is a novel algorithm that combines multiple views of an object to achieve high-resolution 3D surface modeling. The main idea behind RMS is to use information from different angles and distances to reconstruct a more accurate and stable 3D model. The algorithm works by minimizing the reprojection error while accounting for geometric and photometric errors. Applications of RMS are diverse and cover a wide range of industries, including robotics, manufacturing, and medical imaging.
-
Reflectance Models
Surface modeling, an essential engineering technique, involves creating digital representations of surfaces. Reflectance models, a mathematical formula that describes how light interacts with surfaces, are widely used in surface modeling. They can be applied to create virtual tours, remote sensing images, Lidar point clouds, autonomous vehicle navigation systems, and quality control inspections. Reflectance models can be created through manual creation, statistical analysis, or machine learning. These models' wide range of applications make them crucial tools in fields such as engineering, architecture, and geophysics.
-
Realistic 3D Scene Generation
Generating realistic 3D scenes has been a challenge for computer graphics. With advancements in image processing, infrared (IR) and laser scanning, researchers have developed techniques to create more accurate and detailed 3D models of objects and environments. Image-based 3D surface modeling involves capturing an object's image and reconstructing its 3D structure using computer vision algorithms. IR-based 3D surface modeling uses IR sensing technology to capture high-resolution 3D information from the environment. Laser scanning is a non-contact technique that captures detailed information about the surface of an object or environment, producing a digital point cloud for reconstruction. These techniques have applications in manufacturing, architecture, automotive, robotics, navigation, planning, engineering, healthcare, and prosthetics.
-
Real-Time Object Recognition
In this article, we explore a novel approach to real-time object recognition using 3D surface modeling. By combining multiple modalities such as images, infrared, and laser data, we can build more comprehensive representations of objects. Image-based 3D surface modeling techniques generate high-quality 3D models from multiple viewpoints, while infrared data provides insights into the physical properties of objects. Laser scanning technology captures details such as curvature, edges, and holes. Once we have constructed 3D surface models, we can apply various machine learning algorithms for object recognition. This technology has significant potential in applications such as autonomous driving, robotics, and surveillance.