Examples
Typical applications for Wolution
The Institute of Applied Physiology of University of Ulm uses Wolution to quantify differences in immunofluorescence signals in cell images. The key difficulty was to automatically segment cells into their membrane, cytoplasm and nucleus. This is extremely time consuming by hand, but could not previously be automatized. With the Wolution Learning Algorithms this challenging segmentation problem is solved with good fidelity, using a few dozen training images. Wolution also provides the desired intensity measurements and other statistics of the different cell regions in convenient CSV files.
The department of geology of the University of Bayreuth wanted to automatically find dolines in large quantities of satellite images. Dolines are small sink holes that can easily be confused with other structures in the image. A trained geologist can do this by eye, but it is difficult to teach these subtle distinctions to an algorithm. With our newest deep learning segmentation algorithm we were able to get excellent results. Our algorithm provides the coordinates of the dolines in a format that can easily be read in GIS (geographic information system), the standard image and data processing software in geography.
The Bayerische Landesanstalt für Landwirtschaft (LfL) studies plant growth to optimise agricultural processes. For a recent project they needed to find and analyze roots in images with non-uniform light conditions and other difficulties. With the Wolution Learning algorithms, it was possible to solve this problem with high quality.
More examples
Wolution is equipped with a large set of algorithms, and also allows you to train your own algorithms by providing labeled examples. It is therefore possible to perform a wide range of image analysis tasks, and we are continuously working to enlarge the possibilities. Below we list some tasks that you can solve with Wolution.
Segmentation
An important class of image analysis tasks is segmentation. The goal is to select certain regions of the image for further analysis (for example calculating the size of the region, or the number of independent regions). Wolution provides both standard algorithms that can work out of the box, and learning algorithms where you teach the segmentation algorithm by providing hand segmented images, which you can easily generate with our interface. In this example we are finding all neurons in a microscope images of a rat brain with Wolution learning.
Object Detection
Another main class of tasks is object detection. In object detection the goal is to find all objects of a given type in an image. For example you might want to select a class of cells and count them. In this example we are finding cars in satellite images. Programming such a detector by hand would be challenging but with Wolution learning it can be done easily without any image analysis knowledge.
Standard Image Processing
Automatise simple image processing tasks and run them on a large number of images. You might want to change the size or contrast of an image, apply filters, or use denoising algorithms to improve the image quality. Here we are applying a simple edge detector to find the structures in the image.
