Patozone

Colocalization

 

No colocalization def + No stats = No Science

- NOT possible to objectively decide colocalization by eye in RG merge image

- color merge image not good for much

- CCD/PMT (explained below) see intensities different from ppl's eyes

- journal image ink in CMYK for print, screen in RGB

- lossless compression is ok; the image is data!

- a pixel is not a little square!! it is a point

- Point sample = Picture Element = PixEl

- pixels are samples of intensity from a point in space, and the pixel size is pixel spacing distance

- also, a voxel is not a little cube

- what can you digitize? Dimensions!!

 

Detect information loss

 

can split an image c multiple channels into separate windows - Image--> Color --> Split Channels

- zooming in (mag glass and L click; right click zoom out) shows square-shaped artifacts, which is a JPEG lossy artifact (messed up for image analysis)

-- lossy compression should not be used to store scientific data (bad for measurement)

 

Detector saturation / over exposure

clips the saturation peak at the top, wont be able to see how high that mountain is...

- look at the histogram of an image (Analyze--> Histogram; then look at the "Log" button); if there are any pixels at the max or min of a histogram of an image, then there is loss of intensity information

 

Charge-Coupled Device (CCD) - used in high-quality digital imaging; in CCD image sensors, pixels represented by p-doped MOS capacitor, converting photons into electrons at the semiconductor-oxide interface (CMOS active pixel sensors used in digital cameras almost as good quality)

 

Photomultiplier Tubes (PMT) - low-noise, high-sensitive vacuum light detectors in the UV, visible, and near-IR parts of the spectrum

-- Secondary emission is the ability of electrons in a vacuum tube to cause the emission of additional electrons by striking an electrode

 

Incorrect Detector Offset

Can mess up quantitative analysis

- normally, electronics convert light signal to number (0-255 in 8-bit, 0-4095 in 12 bit, or higher!)

- detector offset is the user telling the computer that zero photons are hitting the detector

- a detector offset that is off can be too high or low, and clip off image information or waste the dynamic range of the detector bc lntensity of the digital img not close to linear proportionality c the amt light hitting the detector

- looking at a histogram, too high offset has empty of small numbers at the low end of the intensity scale, meaning there are no or too few pixels c values close to zero

- no easy way of telling if there is too low offset

 

Image analysis

 

Digital images are respresentation of reality, an artifact c less info than the object, just a bunch of numbers, not analog art...

- point spread function (psf)

- images are information c meta data (data about what, where, when and how), noise / background; can be used in quantification

- manipulating an image is changing information, and lost information is lost forever!

 

Shortcuts

 

i = information about an image

shift+z = open channels switch