Tuning vision systems for next-generation cameras

As computational and computer vision cameras continue to overtake conventional cameras in every day products, the difficulty of integrating new optical and software designs is becoming increasingly clear. Tuning ISPs for image quality can be onerous and time-consuming at the best of times, and this is magnified when incorporating innovative optics, sensors and computer vision tasks for every product.

The solution is CRISP-ML:
the Computationally Reconfigurable Image Signal Platform – using Machine Learning




Integrating computer vision tasks into your ISP tuning process is a significant challenge. Whether you are building vision systems that deal with event detection, process control, navigation or automatic inspection, CRISP-ML will provide the ISP optimization required to ensure the corresponding applications will succeed in the field.

CRISP-ML uses numerical optimization techniques found at the core of many machine learning methods (e.g. Deep Learning) to automatically optimize your full imaging and vision system, from ISP to CV. CRISP can effectively combine large real-world CV training data sets together with standards-based metrics to simultaneously improve the performance of CV systems as well as chart driven KPIs. By exploiting the ability of machine learning based methods to leverage large data sets, CRISP-ML automates the tuning steps that are otherwise painful, costly, and time-consuming.

Easily adapts to new optics, sensors, CV tasks and applications.


  • Stereo cameras
  • Array cameras
  • Lenslet arrays
  • Light-field cameras
  • Optical masks
  • Coded-aperture
  • Etc.


  • RGB-IR
  • Foveon X3
  • 3CCD
  • New Bayer filters
  • Multispectral imaging
  • Medical scoping
  • Etc.

         CV Tasks

  • Event detection
  • Process control
  • Navigation
  • Object modelling
  • Automatic inspection
  • Etc.


  • ADAS
  • Security
  • AR/VR
  • IoT
  • Wearables
  • Robotics
  • Mobile
  • Etc.