BiCS student Dany PAIS has contributed to the enhancement of 3 million
images from old newspapers from Luxembourg and Switzerland. The images
were provided by the Luxembourg Centre for Contemporary and Digital
History (C2DH), as part of the Impresso project. The new project,
entitled Spresso, has been conducted under Dany’s BSPs 5 and 6.
Under the supervision of Professor Luis Leiva, Dany worked on his BSP5
to implement a restoration super-resolution based on Deep Learning. As
a result, new synthetic images have been created from the source
material with an often surprising degree of detail. Then, Dany worked
on his BSP6 to apply an intelligent grayscaling algorithm in order to
save valuable storage space, near 50% of the original capacity.
More info: https://wwwen.uni.lu/fstm/news/spresso_historical_newspaper_photography_goes_supercomputing2