A User’s Guide to Assessing Image Quality for Artificial Intelligence | Webinars

Tue, Nov 29, 2022 10:30AM EST

ABOUT THIS WEBINAR

The size and number of images for image acquisition instruments and processing techniques have increased exponentially, along with the associated storage needs and network requirements. Most of these images are processed with machine learning techniques that rely on information invisible to the naked eye, but which can be revealed by observing fine correlations between pixels. Therefore, it is important that the raw images meet specific standards. Currently, however, there is a lack of standardized definitions and quality metrics for these images. Dotphoton works on these issues with associations such as QUAREP-Limi.

Arianne Bercowsky, Ph.D., of Dotphoton, provides an overview of recent work involving normalization and image quality assessment to help future-proof image data. She discusses raw images from a technical and physical point of view, their main characteristics and how to assess image quality to ensure data-centric artificial intelligence (AI) and machine learning. . It will also demonstrate Jetraw’s high-performance raw image compression technology, both as software for the biomedical and pharmaceutical industry, and as a Field Programmable Gate Array (FPGA) implementation. for camera manufacturers. The objective of Jetraw technology is to address big data issues, such as storage space and associated costs, CO₂ emissions and data transfer rates.

Bercowsky is joined by a representative from one of Dotphoton’s partners, Excelitas PCO, who presents a case study on the result of applying Jetraw compression to light sheet microscopes using PCO cameras, demonstrated by Imperial College London.

Who should attend:

Researchers, engineers and manufacturers who use artificial intelligence and machine learning. Laboratory managers, facility IT managers, clinicians, image analysts and camera manufacturers looking to improve image quality assessment. Those interested in or working with light sheet microscopy, drug development and research in industries such as medicine, biomedicine, pharmaceuticals and cancer research.

About the presenters:

Arianne Bercowsky, Ph.D., is an application specialist at Dotphoton. She obtained her doctorate in bioengineering and biotechnology at the École Polytechnique Fédérale de Lausanne (EPFL) in the laboratory of Professor Andrew C. Oates. Oates’ lab was one of the first Dotphoton customers to integrate Jetraw technology into their image acquisition and processing workflow. The lab’s accelerated datasets were measured in terabytes, which made data management, storage, and transfer problematic. Oates’ lab cut data transfer time from two hours per data set to just 15 minutes, reducing costs and the lab’s carbon footprint along the way. Bercowsky saw an immediate need for such technology in biomedical, medical imaging, and machine learning applications. She is now passionate about helping researchers improve their image data acquisition and processing workflow.

About Dotphoton:

Dotphoton is a Swiss software company providing scalable solutions for processing large image data. Its flagship product, Jetraw, is the first raw compression technology perfectly suited for data-centric artificial intelligence (AI) and machine learning. Jetraw is delivered as both software and a Field Programmable Gate Array (FPGA) and improves the management of large data sets for optical system manufacturers and their end users. Dotphoton’s metrologically correct compression reduces file size, speeds up data transmission by at least 4 times and also reduces storage costs and carbon footprint. This allows companies to meet their performance and environmental goals without sacrificing data quality. Dotphoton’s partners and customers include the European Space Agency, Bosch, leading camera manufacturers for life sciences and biomedical laboratories.