Image analysis refers to using a computer to automatically take measurements on an image (usually many many images). Almost any feature or quality that can be imagined, such as number, size, shape, or intensity, can potentially be measured in an automated and objective way. When the automatic routine is not accurate enough, limited user intervention can be used to ‘clean up’ the final result. We offer full support for ImageJ/FIJI, Imaris, and Ilastik. We can also work in Matlab for large or especially challenging projects, even on a collaborative basis. Please contact us with questions.
Channel Specific Segmentation
Most simply, the ‘objects’ on each monochrome channel can be identified and measured. In this example, all the nuclei (blue) within a particular tissue region (red) are being counted automatically.
In some cases, the ‘objects’ of interest are defined based on the staining across multiple channels. These cases are more challenging, but can usually still be largely automated and then refined using minor user intervention. In this example, specific types of nerve terminals are identified based on two channels as well as the shape of the unstained regions that the terminal encloses.
EM images are particularly challenging to segment because the objects of interest are defined only by edges, rather than by intensities. Still, moderately automated segmentation is possible. This example shows an automatic segmentation of membranes (red) and mitochondria (blue) in a slice of a FIB-SEM data set.