Fusion of plant histology information obtained from different imaging modalities
Crop species such as maize are of growing interest for cattle feeding or biofuel production.
Their valuation depends mainly on the chemical composition, on the cellular morphology, and on the
spatial organization of different tissues within plant stems. The histology of stem sections can be
investigated using various imaging modalities that provide different information:
microscopy and macroscopy imaging have resolution that makes it possible to quantify the cellular
morphology; histochemical staining provides partial information on chemical composition, at a lower
resolution; fluorescence imaging provides insight of local chemical composition, but requires
statistical processing and modelling of the data.
We aim at intregrating and merging the information obtained from different imaging modalities on
various collections of samples. Coupling of image information involves several processing step
including the extraction of meaningful information from the different imaging modalities,
the image registration of individual slices observed with different modalities,
the group-wise registration of images of slices with various shapes and sizes to create a statistical atlas, and the statistical analysis of descriptive features.
We will present strategies for different collections of images, obtained from different modalities,
and with various experimental conditions.