Philips Healthcare MRI Image Reconstruction

Philips Healthcare Uses the Intel® Distribution of OpenVINO™ Toolkit and the Intel® DevCloud for the Edge to Accelerate Compressed Sensing Image Reconstruction Algorithms for MRI. The Intel Distribution of OpenVINO toolkit allows developers to deploy their deep learning models with improved inference on a variety of Intel hardware. Using the toolkit’s custom extensions feature, Intel was able to speed up the compressed sensing workloads for Philips Healthcare by as much as 54x on an Intel Xeon Gold 6138 processor compared with unoptimized TensorFlow. Philips Healthcare was also able to leverage the Intel DevCloud for the Edge to quickly benchmark their CS models on Intel CPU, VPU, FPGA, and integrated GPU hardware. This platform helps Philips Healthcare when they plan and develop new products by allowing them to easily assess various deep learning pipeline parameters such as performance, price, power, and form factors for their designs.