PX3D-AI Core Product

Photon-X's goal of Artificial Intelligence (AI) is to provide machines the capabilities to learn, think, and reason like humans or in some cases, better than humans. To achieve these goals, the input data must exceed human cognitive-sensory abilities to create and train a truly self-adaptive artificially intelligent system. An AI system should: Identify, Qualify, and Quantify in real-time with high fidelity.

The Photon-XI 3D Shape AI technology product enhances current 2D AI capabilities with passive high-resolution optical 3D surface data and analytics  necessary to achieve human-like AI. 

3D Shape-based AI Advantages

Photon-X AI Data Advantages: Proprietary and Patented: 3D angle data at the pixel level – a significantly enhanced optical dataset.

 

  • Recognize objects in passively in real-time with a single sensor

  • Angles are an invariant feature across objects and EM spectrum

  • Directionality of all light sources (global light map)

  • Correlate by angle, rates of slope, edges, slope, rates of slope, objects, and sub-sections of objects

  • Segment by angles, pieces and parts of objects or scenes

  • Analyze by object – not points; drill down to sub-units

  • Object orientations and intensity do not matter

  • Calculate the index of refraction to determine material composition

  • Measure surface features

  • Correlate multi-sensor modality – slope vector analysis

  • Point clouds unnecessary

  • Rendering required to provide solution

  • There are many more advantages over scalar 2D AI

Key PX3D-AI Core Product Aspects

 

Spatial Phase Imaging

Spatial Phase Imaging (SPI). The core technology is a single proprietary passive sensor that captures shape data, outputs pixel level 3D surfaces, works on dynamic objects, and across the EM spectrum - all in real-time. SPI provides high resolution images, shortens the post sensor processing by not requiring point clouds, and ads a new dataset to Deep Neural Networks and AI. The new dataset helps eliminate the nil conundrum, digital spoofing, and other color/intensity data analysis issues. SPI works on land, sea, air, and space.

First Order Applications

As the SPI sensor outputs a new dataset, Photon-X has created and continues to create shape algorithms to solve clients' problems. In the long-term, Photon-X plans to provide a developer’s toolkit (SDK) to allow third parties to build the analytics and AI algorithms utilizing the proprietary 3D surface data sets. 

Artificial 3D AI

Synthetic 3D Shape AI utilizes older multi-frame data sets and reprocess to build 3D shapes for AI training.

 

  • Synthetic AI can ingest and process very large datasets

  • The 3D Shape AI technology is backwards compatible

  • Change the Scene Environment for training, including:

    • Additional angles of light

    • Alter the type of light inputs (e.g. frequency & color) 

    • Alter the specular and diffuse components of the surface

    • Material types

    • Time of year

    • Position on earth or space (e.g. desert, mountains, forest)

    • Color 

  • Objective is to Build better AI training datasets for the SPI sensors whereby objects can be recognized by SHAPE

3D Shape-based AI

 

1.The angle relationship between the top and side and the windshield of the Humvee never changes. This geometric relationship remains the same no matter the position of the viewer.

2.As the target and camera move relative to each other, subsequent frames add more data and allow newly visible panels to be built and appended the object.

© 2020 Photon-X, Inc 

  • wikipedia-icon-png-5.jpg
  • YouTube
  • LinkedIn