The fundamentals of our algorithmic approach are straightforward, proven, and patented. It is based on the principles that:
- Materials respond to activating (electromagnetic or acoustic) energies deterministically and that the response can be accurately quantified and classified.
- Pixel data representing an object, or material type, encompasses a coherent set of properties or pixel values that represent relationships or “signatures.”
- Signatures provide a significantly more relevant set of features than what can be obtained from single feature measurements such as density or Zeff.
The core analytical process utilizes a series of iterative transformations and convolutions that can be applied sequentially and or in parallel. In these sequences the output of one process is the source of the next process.
For example, the two different materials in the original x-ray image, shown on the left, appear visually indistinguishable. After applying an iterative algorithmic sequence to the original image, the color image below exhibits the differentiation that actually existed at the micro-contrast level. After processing, the differences in properties between the Pyrodex and water become visually apparent.