Multimedia, Image and Video Processing, Embedded products, Portable products, Algorithms and mathematics, Processor architectures, RTL and semiconductor IP, Software engineering, Product definition, Business advantage
Imsense is a software start-up focused on improving images through dynamic range correction. Their proprietary algorithms had been developed for use on a standard PC computing platform. The challenge was for Argon Design to make use of our in-depth graphics, imaging and semiconductor knowledge to determine whether it would be practical to implement Imsense’s algorithms in hardware, making them suitable for use in handheld and video applications.
This was a strategic opportunity for Imsense, who were pursuing two Tier 1 OEMs requiring hardware implementations of their software. Imsense having no hardware IP design experience in-house, selected Argon Design because of our extensive experience in semiconductor IP, chip design, system design and stills/video imaging and pipelines.
The project was addressed in two phases, first to analyse the existing algorithms and explore alternative options, before creating the optimal architecture design to implement the final algorithm.
During phase one we examined a number of options alongside the original Imsense algorithm in order to understand the advantages and disadvantages of each approach. The key issues being speed of execution, computing/memory requirements, size and estimated power consumption as the final solution would be required to run in real-time on a mobile device. Trade-offs had to be made and a deliverable quality levels agreed with Imsense, along with methods to protect the original images.
Phase two took the output from phase one and led to us producing overview proposals on how best to implement the algorithms in a mobile solution. The initial proposals included IP requirements in terms of computing power, memory and alternative processing methodologies, as well as producing IP size estimates in terms of estimated gate count. The proposals would then be extended to include hardware IP design, firmware development and control software once Imsense had secured their lead customer.
The output from the project provided significant detail on how the algorithms could be implemented, along with additional ideas for Imsense to further optimise their algorithms for mobile use. The follow-on stage of actual IP development was not pursued as Imsense was acquired by Apple, which took the project in-house.
The split image below shows the contrast and quality using Dynamic Range Correction (DRC). On the left hand side DRC has been applied to the image, on the right hand side the image remains un-processed.
Do you have a project that you would like to discuss with us? Or have a general enquiry? Please feel free to contact usContact us