This company develops a high-performance Visual-Inertial SLAM solution and an AI-based Visual Positioning System (VPS) implemented in an SDK. Their technology fuses data from cameras and IMUs to compute accurate device location in real time on embedded CPUs. The SDK supports various sensors and processors for applications in robotics, VR, and computer vision research.
Funding
Funding not disclosed
Founders
Product
Problem
Accurate real-time localization and spatial understanding are critical for advanced robotics, virtual reality, and computer vision applications. Traditional methods often struggle with sensor fusion, computational efficiency on embedded systems, and robustness in dynamic environments.
Solution
Spectacular AI provides a high-performance Visual-Inertial SLAM (VISLAM) solution that accurately tracks device location by fusing camera and IMU data. Their proprietary VISLAM algorithm, an evolution of state-of-the-art methods like HybVIO, offers robust and precise pose estimation. The system is designed to run efficiently in real-time on embedded CPUs, supporting a wide range of sensors and processors for flexible integration. This technology enables applications requiring precise spatial understanding and real-time mapping capabilities.
Target Audience
The primary target audience includes developers and researchers in robotics, augmented reality, virtual reality, and computer vision who require accurate, real-time spatial tracking and mapping capabilities.
Features
- Proprietary Visual-Inertial SLAM (VISLAM) algorithm, building upon research such as HybVIO for enhanced performance.
- Real-time operation on embedded CPUs, ensuring efficiency for edge devices.
- Fusion of camera and Inertial Measurement Unit (IMU) data for accurate pose estimation.
- Support for a wide range of sensors and processors, facilitating hardware flexibility.
- Optimized for applications requiring precise spatial understanding and real-time tracking.
- Publicly available SDK wrappers for rapid prototyping with common devices like OAK-D, RealSense, Azure Kinect, and ORBBEC.
- Research contributions include advancements in Gaussian Splatting for motion blur and rolling shutter compensation.