CRISP: The Image Signal Processing pipeline for next-generation cameras

As major disruption to the conventional camera comes from computational cameras, new optical and software designs are not easily integrated into the existing ecosystem. Furthermore, the existing silicon systems are not powerful or power-efficient enough to support new designs.

Enter CRISP, Computationally Reconfigurable Image Signal Platform

Conventional camera pipeline

Conventional

Separate processes aggregate errors

  • In conventional cameras, each imaging process is separately integrated to optics and hardware, which vary across devices.
  • Each process creates errors that carry over to the next process in the pipeline. These errors are aggregated, creating suboptimal image & use of resources. For OEMs, this means a painstaking integration and testing, often delaying production.

CRISP pipeline

CRISP

A unified framework optimizes for quality

  • CRISP is an image reconstruction engine that incorporates all processes into one unified framework.
  • CRISP works efficiently between any existing optics and any hardware, and its objective is to optimize final image quality.

For OEMs, CRISP means:

Higher quality images, Simpler integration, Better use of resources


CRISP v1

Virtual Lens

Virtual Lens corrects optical aberrations. This enables sharper, more detailed photos. Additionally, Virtual Lens empowers flexibility in lens design that can be used to reduce the cost and/or the number of lenses. This approach can also be used to improve camera production yield.

Examples of Virtual Lens – deblurring a picture taken with a smartphone’s camera:

 
  • Before-d
    After-d
    Image capture d With Virtual Lens

Mobile camera, 8-megapixel sensor (Image cropped to visualize details)


  • Before-d
    After-d
    Image Capture d With Virtual Lens

Mobile camera with damaged lens, 8-megapixel sensor (Image cropped to visualize details)


  • Before-ccc
    After-ccc
    Image Capture (damaged lens) ccc With Virtual Lens

Mobile camera with damaged lens, 8-megapixel sensor (Image cropped to visualize details)


  • Before-Wood
    After-Wood
    Image Capture Wood With Virtual Lens

Mobile camera with damaged lens, 8-megapixel sensor (Image cropped to visualize details)


Benefits


BetterPictures-(2)
Devices can capture clearer pictures with existing optics.


ThinnerCameras

Flexible lens design can be used to reduce the weight and z-height of a camera module, making for a thinner device.


CheaperOptics-(2)The quantity and quality of the optical elements needed is diminished, allowing manufacturers to obtain similar or better results at a lower cost.


Higher-Yield300x

Correcting optical aberrations in camera modules leads to higher manufacturing yields.


Algolux Virtual IS

Algolux Virtual IS provides image stabilization by correcting motion blur and shutter shake, two common problems that arise when taking pictures in low-light conditions. Virtual IS uses a secondary camera (for example, the front-facing camera on a smartphone) to track the micro-movements causing the blur. It then uses this information to deblur the image. The resulting low-light pictures are sharp and blur-free.

Example of Virtual IS – deblur a picture taken with a smartphone’s camera:

  • Before-VIS
    After-VIS
    Image Capture VIS With Virtual IS