TechFinity's
image processing toolset is a set of algorithms based on the
active contours method of segmentation, which uses region-based
rather than edge-based techniques. This permits fast and
robust object extraction, registration, and tracking in
image data of varying quality. The region-based segmentation
technique is less sensitive to noise allowing robust
segmentation to be performed without denoising or other
preprocessing.
The
region-based techniques can be used for real-time
object tracking applications without requiring
preprocessing for noisy or low intensity data.
Simultaneous
multi-band image processing is accomplished using
standard combining techniques as well as innovative
logic-based modeling. Simultaneous segmentation
of the same image at different bandwidths generates
a fused image where the multi-band objects are
combined according to user-specified logic rules
(e.g. union, intersection, etc.). This fused image
can often provide information that is not easily
obtained from the individual images. The union of
the objects in the images, the intersection of
the objects, and identification of objects in one
image that are not in the other image, make our
image processing toolset applicable to a wide domain.
Multi-image
source panoramic view algorithm can stitch together multiple
camera views that have a certain degree of overlap, and create
a single integrated panoramic picture. The algorithm
incorporates local image registration to fine-tune the
overlap in the final, panoramic image.
