|
Image
Analysis Tools
ImageQ™
and DiscoverQ™ provide scientists with the capability for easier
HCS assay development, which is one of the critical bottlenecks
in using cellular imaging technology in drug discovery.
ImageQ™
Cytoprint's proprietary "shotgun" image analysis approach
applies a large number of image feature extraction operators to
create a descriptor vector, or Cytoprint™, of the image changes.
These vectors are statistically analyzed to determine the features
that are relevant for a given assay task. This system, ImageQ™,
contains the individual image operators in the shotgun application.
There are currently several hundred image operators and Cytoprint
is continually adding more operators to our ensemble. Scientists
analyzing cellular images can use this system to rapidly apply and
evaluate image analysis techniques they select from a large library
of algorithms that have been developed and optimized for cellular
imaging.
DiscoverQ™
Statistical analysis of image vectors by DiscoverQ™ is Cytoprint's
core technology for converting HCS imagery into a numerical, dataminable
signature vectors. Proprietary informatics tools in DiscoverQ™
automatically discover image-based assays from the feature extraction
algorithms in ImageQ™. Scientists can use the feature selection
methods implemented in DiscoverQ™ to select the minimal set
of image features that are statistically relevant for a given assay
that distinguishes between two given cell phenotypes. The resulting
selected image processing algorithms, along with the cell lines,
markers and imaging protocols used to collect the data then form
the basis of this HCS assay. Cytoprint currently has patents pending
in this area of assay discovery.
The numerical
methods in DiscoverQ™ also enable the statistical comparison
of cellular images through their biologically-relevant features,
i.e., via phenotype differences present in the imagery and manifest
in the features extracted with ImageQ™.
|
