Question: Looking for a solution that can perform object recognition and motion capture from a collection of photos or video frames.

If you need a system that can recognize objects and track motion from a group of photos or video frames, Move AI is a good option. It combines AI, computer vision, biomechanics and physics to convert 2D video into 3D motion data. The system works with single and multacameras and can include features like hand and finger tracking, high-quality motion data and retargeting to common 3D animation software formats. It's designed to scale and has been used in games and advertising.

Another option is Movella, a full-stack motion capture technology that can be used in a variety of industries. It offers high-quality motion capture data, precise location and orientation tracking and ergonomic analysis for safer working conditions. Movella's products include Xsens Motion Capture, Xsens Animate, Xsens Analyze and Xsens Link, and it can be integrated with common 3D software packages like Unity, Unreal and Autodesk.

For more photogrammetric uses, AliceVision can help you create detailed, accurate 3D models from photos or video frames. It's good for scene modeling, object recognition and motion capture, and can be used in movies, architecture and visual effects. The technology can be used to reconstruct 3D scenes and reconstruct motion from 2D images.