Instructor: Bilal Zafar and Sohaib Imran, 10xEngineers
Level: Intermediate
Prerequisites: Bayer Image Sensors, image processing basics, rudimentary knowledge of digital camera pipelines; programming: Python / C.

Benefits:
This course enables the attendee to:

• Build knowledge base on hardware ISP development cycle and its components (algorithm exploration, Golden Model, RTL ISP, FPGA / ASIC implementation, ISP firmware, ISP tuning tool)
• Gain hands-on experience in porting a new image sensor to an ISP (ISP tuning)
• Gain hands-on experience in tuning ISP to achieve better output image quality)
• Handle Video Frames Processing, 3A algorithm tuning with Infinite-ISP
• Process an available dataset with Infinite-ISP for your vision or AI / deep learning application
• Collect dataset of burst capture uncompressed sensor RAW and ISP output for your vision or AI work

Course Description:
The course takes students through the open-source hardware image signal processor development suite, Infinite-ISP. Users learn the different components of the Infinite-ISP package, their roles in the hardware ISP development cycle and the underlying design methodology to develop an image signal processor from scratch. The course enables users to port a new image sensor to an ISP and tuning the ISP algorithm blocks for improving the output image quality. It discusses some other developer-friendly features of the Infinite-ISP software pipelines such as video processing and 3A rendering. The course finally throws light on useful applications such as custom dataset capture with Infinite-ISP as well as AI model / vision pipeline performance enhancement with ISP tuning and optimized image pipeline configuration.

Intended Audience: AI researchers, vision SoC architects, imaging & CV experts:

• AI researchers: can use this course to capture and access royalty-free uncompressed sensor RAW + ISP output of tunable image quality for their models & vision pipelines
• Vision SoC architects: can use this course to develop custom ISP for their application and learn to integrate open-source ISP into their systems
• Imaging & CV experts: can use this course to design custom camera ISP algorithms and pipelines as well as design AI-assisted algorithm blocks while gauging the hardware resource utilization

Bilal Zafar is the co-founder and CEO of 10xEngineers, a hardware design services company specializing in RISC-V microprocessors and camera image-signal processors. Previously, Zafar was principal engineer at Qualcomm, where he co-led a world-wide team working on developing custom and semi-custom IPs. He holds an MS and PhD in computer engineering from the University of Southern California, Los Angeles, and a BSc in electronics engineering from the GIK Institute of Engineering Sciences & Technology.

Sohaib Imran Bhatti is a Senior Camera ISP Algorithms Engineer at 10xEngineers, California. He received his bachelor's degree in electrical engineering from the University of Engineering and Technology, Lahore in 2017. Since then he has worked on diverse projects in the IT industry. Over the past five years, he has specialized in Camera Image Signal Processors (ISP), beginning his career as an Algorithms Engineer, where he developed cutting-edge calibration algorithms for commercially available ISP tuning tools. He later led the end-to-end development of Infinite ISP, an open-source camera solution. As the technical lead for Infinite ISP, he was responsible for overseeing the entire project lifecycle, from concept to execution, while playing a key role in system architecture and leading a team to address complex imaging challenges. His experience includes collaborating closely with hardware teams to ensure seamless integration of the ISP with FPGA platforms, including Efinix and Xilinx Kria. Additionally, he has expertise in camera color correction and lens shading correction algorithms.