Bioscience Imaging Facility
Custom Imaging Processing
Digital images can be analyzed qualitatively (e.g. by looking at them) or quantitatively. In either case, some sort of digital imaging processing is often desirable or required prior to analysis. Of course, the affects of all processing steps must be fully understood and described, in order to ensure that an image's scientific integrity is maintained. Often, images may also need to be resized prior to publication.
In order to facilitate the proper use of these specialized image processing techniques, the Bioscience Imaging Facility provides an imaging processing service for a flat, $10 per image fee, which includes a short 'methods' section describing the processing steps applied.
Three broad classes of image processing techniques are available: 1) Contrast and color adjustments, 2) Spatial filtering, and 3) Image resizing. An example of each of these techniques is shown below. Advanced techniques (e.g. drift correction, extended depth of field, deconvolution, co-registration) are also available - Please inquire.
Contrast and color adjustments
- (A) An unprocessed image of a fixed, mammalian cell nucleus during mitosis (60x, Nikon C1+).
- (B) The same image after application of linear contrast adjustments.
- (C) Application of non-linear contrast adjustments. Non-linear adjustments preclude quantitative intensity measurements, but are useful for accentuating fine (dim) structural details.
- (A) Signal represents the bio-distribution of a radio-lableled tracer within the body of a living mouse immediately after acquisition and reconstruction (MiLabs SPECT/CT). Note the presence of substantial noise. (Noisy images are also encountered in microscopy when signal is very weak, as occurs during high-speed image acquisition).
- (B) Judicious smoothing techniques remove noise, while maintaining detail, thereby increasing lung tumor visibility.
- (C) The smoothed SPECT signal overlaid on the corresponding CT image.
- (A) A group of living yeast cells undergoing cell-division (60x, Nikon A1Rsi). Since these cells are very small and many cells were imaged per field, sub-cellular detail was 'under-sampled', meaning that not enough points were imaged to adequately represent finer details.
- (B) Consequently, standard digital zoom of the boxed region in A results in a strongly pixelated image.
- (C) Specialized enlargement techniques can greatly reduce pixelation, although, of course, no new image detail can be realized.