Funding
Funding not disclosed
Founders
Product
Problem
Accurate motion analysis for humans, vehicles, and animals requires high‑precision pose estimation, but existing solutions often need specialized hardware, run at low frame rates, or struggle with occlusions and multi‑species scenarios, limiting their usefulness in sports, autonomous driving, healthcare, and animal research.
Solution
Dianar offers an AI‑powered pose estimation platform that delivers real‑time, high‑accuracy tracking of humans, vehicles, and animals across diverse environments. Its deep learning models extract detailed key‑point and orientation data at 60+ frames per second on standard hardware, enabling immediate analysis without costly infrastructure. The service provides a cross‑platform API and edge deployment support, allowing developers to integrate pose data into applications such as sports performance tools, autonomous vehicle systems, and behavioral research pipelines. Advanced attention‑based feature extraction, temporal consistency, and occlusion handling ensure reliable tracking even in complex scenes. Users can access the platform via a simple Python SDK, retrieve confidence scores, and feed the results into downstream analytics or monitoring workflows.
Target Audience
Primary customers include sports analytics firms, autonomous vehicle developers, healthcare and rehabilitation providers, and researchers conducting animal behavior studies who require fast, accurate pose data for their applications.
Features
- 17‑point skeletal tracking for human biomechanics with medical‑grade precision
- Vehicle pose estimation capturing orientation, velocity vectors, and structural components
- Multi‑species animal pose tracking with support for various species categories
- Real‑time inference at 60+ FPS on standard CPUs/GPUs
- Attention‑based multi‑scale feature extraction and temporal consistency algorithms for smooth tracking
- Edge deployment with optimized inference engines for low‑latency applications
- Comprehensive API ecosystem and Python SDK for easy integration
- Advanced occlusion handling and partial pose reconstruction