“Image” refers not only to RGB photographs, but to any pattern that
describes the features or state of an object or phenomenon.
There are diverse forms of images representing various aspects of the
real world—such as depth images, spectral images, CT and MR images. In
mathematics, “image” also means the set resulting from mapping a
structure from one space to another via a function.
In “Image Informatics”, we comprehensively study the structures,
transformations, and representations of such “images,” which arise in
various descriptive spaces through sensors and computational processes,
aiming to approach the fundamental essence of the real world.
– Computational Imaging: An image is not the object itself, but a
modality-dependent representation of certain aspects, determined by the
method and conditions of observation. Observation can be regarded as
“encoding,” while extracting fundamental patterns and structures can be
seen as “decoding.” This process involves technologies such as optical
design, electronics, machine learning, and optimization.
– Understanding Motion and Deformation: Temporal changes in images
are key to understanding and predicting phenomena. We use geometric and
statistical approaches to model diverse changes—such as motion
constrained by skeletal structure, or the dynamic deformation of clouds
and biological tissues—aiming for unified analysis of dynamic “images”
in the real world.
– Visualization and Expression: Effective presentation requires
careful design of “perspective.” By selecting and expressing only the
most relevant aspects as the optimal “image,” we support discovery and
better decision-making. We explore new ways of delivering images—not
only through 2D displays, but also using 3D displays, 3D printers, and
other modalities.
Image Informatics Laboratory（URL）
funatomi.takuya.2c (AT) kyoto-u.ac.jp
Professor: Takuya FUNATOMI
