Thanks to technological developments, images have become very common data, present and accessible through many different media (TV, camera, smartphone, tablet, and so on). Consequently, it is more and more common for the scientific community to use images to characterize materials or objects under study at different scales. Cultural Heritage (CH) is one of the areas where image and data processing is more relevant to document, understand and share the knowledge associated with it. The introduction of the methods developed in the image-processing domain offers new possibilities to the interdisciplinary CH world.
This Special Session on Image Processing for Cultural Heritage aims to present studies that exploit images as principal data to solve heritage problems. Topics of this special session may include, but are not limited to:
Digitizing Europe's cultural heritage is an active part of the European Union agenda, and the EU has now initiated European Year of Cultural Heritage 2018, with the aim to encourage more people to discover and engage with Europe's cultural heritage, and to reinforce a sense of belonging to a common European space.
EUVIP'2018 welcomes research groups from various countries to submit work on topics of image processing related to heritage, also specialists in cultural heritage are welcome to present their image-processing problems.
During the recent years, breakthroughs in machine learning have led to human level accuracy in a number of problem domains, including image recognition, image segmentation, audio and speech recognition. The disruption was largely due to discoveries in deep learning, with structures such as convnets or recurrent nets, and today their use is ubiquitous and several competing platforms exist for easy training and testing of deep networks.
Most modern machine learning research is devoted to improving the accuracy of prediction. However, less attention is paid to deployment of machine learning systems. Most deployments are in the cloud, with abundant resources, and free choice of computation platform. However, in the advent of intelligent physical devices—such as intelligent robots or autonomous vehicles—there is a need for low-latency implementations running on low-resource edge computing hardware.
To address these questions, we organize a special session during EUVIP 2018 (November 26-28, Tampere, Finland) on deep learning implementations; including both system level topics and other research questions related to general deployment of machine learning algorithms.
This special session seeks innovative new solutions for machine learning with strong implementation and deployment aspect. The list of possible topics includes:
The papers should be submitted through the regular submission system and will undergo the same review process as regular papers.