Summary
Over the past few years hyperspectral (HS) imaging has been broadly applied in a wealth of different applications with
remote sensing of the environment being the most prominent one. HS imaging provides a rich amount of information by
generating images and videos of high spectral resolution captured at a wide range of the electro-magnetic spectrum.
Recently, HS data have been shown to offer remarkable advances to a new field of significant interest i.e., medical HS
(mHS) imaging. The high spectral resolution of HS data makes them amenable to identifying even subtle spectral differences
related to various pathological conditions. In view of that, mHS images and videos have received considerable attention
lately. mHS data have already been used for non-invasive diagnosis of several types of cancer e.g. brain, tongue cancer, as
well as for diabetic foot diagnosis and surgical guidance. mHS imaging is anticipated to remarkably flourish in the years to
come taking into account the recent advances that have occurred in the development of micro-size and low-cost HS
cameras. However, despite this large progress in HS imaging hardware, sophisticated algorithms capable to interpret these
data are still missing. HyPPOCRATES aims at deriving new powerful mHS image and video interpretation schemes tailored
to mHS data processing, by applying novel machine learning ideas. To this end, the problems of subspace clustering and
unmixing will be investigated for performing refined mHS image and video understanding. Along those lines, constrained
matrix and tensor factorization approaches will be explored for devising computationally efficient and scalable machine
learning algorithms. Overall, the main objective of the project is to bridge the gap between the recent advances in mHS
imaging and those in machine learning research. This way, the researcher aspires to go the diagnostic process of several
serious diseases, such as various types of cancer, one step further.
remote sensing of the environment being the most prominent one. HS imaging provides a rich amount of information by
generating images and videos of high spectral resolution captured at a wide range of the electro-magnetic spectrum.
Recently, HS data have been shown to offer remarkable advances to a new field of significant interest i.e., medical HS
(mHS) imaging. The high spectral resolution of HS data makes them amenable to identifying even subtle spectral differences
related to various pathological conditions. In view of that, mHS images and videos have received considerable attention
lately. mHS data have already been used for non-invasive diagnosis of several types of cancer e.g. brain, tongue cancer, as
well as for diabetic foot diagnosis and surgical guidance. mHS imaging is anticipated to remarkably flourish in the years to
come taking into account the recent advances that have occurred in the development of micro-size and low-cost HS
cameras. However, despite this large progress in HS imaging hardware, sophisticated algorithms capable to interpret these
data are still missing. HyPPOCRATES aims at deriving new powerful mHS image and video interpretation schemes tailored
to mHS data processing, by applying novel machine learning ideas. To this end, the problems of subspace clustering and
unmixing will be investigated for performing refined mHS image and video understanding. Along those lines, constrained
matrix and tensor factorization approaches will be explored for devising computationally efficient and scalable machine
learning algorithms. Overall, the main objective of the project is to bridge the gap between the recent advances in mHS
imaging and those in machine learning research. This way, the researcher aspires to go the diagnostic process of several
serious diseases, such as various types of cancer, one step further.
Unfold all
/
Fold all
More information & hyperlinks
| Web resources: | https://cordis.europa.eu/project/id/844290 |
| Start date: | 07-10-2019 |
| End date: | 06-10-2022 |
| Total budget - Public funding: | 259 808,64 Euro - 259 808,00 Euro |
Cordis data
Original description
Over the past few years hyperspectral (HS) imaging has been broadly applied in a wealth of different applications withremote sensing of the environment being the most prominent one. HS imaging provides a rich amount of information by
generating images and videos of high spectral resolution captured at a wide range of the electro-magnetic spectrum.
Recently, HS data have been shown to offer remarkable advances to a new field of significant interest i.e., medical HS
(mHS) imaging. The high spectral resolution of HS data makes them amenable to identifying even subtle spectral differences
related to various pathological conditions. In view of that, mHS images and videos have received considerable attention
lately. mHS data have already been used for non-invasive diagnosis of several types of cancer e.g. brain, tongue cancer, as
well as for diabetic foot diagnosis and surgical guidance. mHS imaging is anticipated to remarkably flourish in the years to
come taking into account the recent advances that have occurred in the development of micro-size and low-cost HS
cameras. However, despite this large progress in HS imaging hardware, sophisticated algorithms capable to interpret these
data are still missing. HyPPOCRATES aims at deriving new powerful mHS image and video interpretation schemes tailored
to mHS data processing, by applying novel machine learning ideas. To this end, the problems of subspace clustering and
unmixing will be investigated for performing refined mHS image and video understanding. Along those lines, constrained
matrix and tensor factorization approaches will be explored for devising computationally efficient and scalable machine
learning algorithms. Overall, the main objective of the project is to bridge the gap between the recent advances in mHS
imaging and those in machine learning research. This way, the researcher aspires to go the diagnostic process of several
serious diseases, such as various types of cancer, one step further.
Status
CLOSEDCall topic
MSCA-IF-2018Update Date
28-04-2024
Images
No images available.
Geographical location(s)
Search
