Just to help you navigate some of the unfamiliar image analysis lingo you encounter:
Image – 2D Array of gray values, with 8bit (0-255 grey values), 16bit (0-65536), or 32bit (6 decimal resolution any number)
Stack – Multiple images stacked on top of each other. Usually makes a 3D data set. 3rd dimension = time, space (depth), color
Slice – one image within a stack.
Montage – A 2D array of images, usually used for visualization and comparison purposes.
Hyperstack – 4D+ images, usually 2D (x-y image space), then time, z position, color, that’s usually it for microscopy data.
Pixel – unit of a 2D image. Each image is composed of a bunch of pixels of 8, 16, or 32 bit depth.
Voxel – same as a pixel, but for 3 dimensional images (usually spatial).
LUTs – Look up tables, or colormaps to false color your images
Autoscaling – helps you visualize the dynamic range of your data better.
Dynamic range – the range of pixel grey values between your maximum and minimum pixel values.
Scale – the size of your pixels in your image, spatially. Usually a few microns or 100s of nanometers.
Projection – collapsing the 3rd dimension of your image to view it’s contents in 2D. (max intentisty, stdev, min, avg, across all images.)
ROI – Region of interest, a selection you make on an image, usually a cell, boundary, or something like that.
Circle, rectangle, polygon, point, oval selections, elipses. EXTREMELY IMPORTANT
Overlay – We usually use overlays for informational purposes. E.g. relative time, depth, scale bar for reference
Stitching – Putting together a bunch of images of a sample that overlap a bit into a larger image.
Import – loading a file into FIJI
Export – Saving a file into a different file format than you wanf in FIJI
Smooth – Useful for evening out noisy images or hot pixels. Reduces edge contrast. Aka Gaussian filter / blur
Sharpen – enhance edge contrast. Fine for visualization or generating masks. Not ideal for quantification.
Noise – variations in the background when your imaging.
Signal – The difference between your highest pixel values from your background.
Filter – used to enhance or suppress features of interest in an image to better extract information out of them
Binary – images where pixels are one of two possible values. Usually 0 and 255, or 0 and 1. Nothing in between. Useful for making masks.
Mask – an binary interest highlighting regions of an image you want to extract or explore.
Feature – Something unique in your image that you are interested in. A peak, the width of a peak, an edge, a line, etc.
Segmentation – a type of classification for images, assigning some type of feature in your image a qualitative value
Registration – Aligning images of similar regions of your sample.
Transform – shifting the pixels in your images spatially in multiple dimensions.
Binary Functions:
Make Binary – pegs all of your pixel values in your image to either 0 or 255, somewhat arbitrarily.
Threshold – makes a binary image (mask) based on specific thresholding parameters you set (e.g. ranges of pixel values)
It is often helpful to combine filters with thresholding to get cleaner masks to operate with.
Erode – Remove outermost layer of pixels @ 255 value in your binary image. Useful for getting rid of noise and hot pixels
Dilate – Adds a layer of pixels along 255 blobs in your image.
Open / Close – lets you generate solid mask objects from noisey images. Can be very useful.
Outline – Does just what it sounds. Turns blob masks into outlines of blob masks.
Fill holes – pretty self explanatory
Skeletonize – collapses all of your binary blobs into a 1px wide object. Super helpful for looking at vessels or dendritic cells
Math – Lets you add/subtract/multiply/divide values from your images
Image Calculator – lets you do math between two entire images. Super useful for normalizing.
Calculator Plus – See above. Useful for different reasons
FFT – Fast Fourier Transform. Generates a map of spatial frequencies in your img (everything is made of added sine waves)
Measure – Gives you image statistics about your image, stacks, ROIs within those imgs. Modify in Set Measurements.
Analyze Particles – SUPER USEFUL FOR GENERATING ROIs FROM MASKS!!!
ROI manager – Lets you list a bunch of ROIs and save them for later.
Classify – useful for assigning qualitative descriptions to image data, e.g. where’s my cells, where’s my background, etc.
Colocalization – how well to intensities of two different images overlap. May need to read up on details.
Plot Profile – measures pixel values from images / ROIs in the 3rd dimension of your image.
Histogram – the distribution of pixel values in your image. Sometimes useful for assessing / classifying images.
Macro – a set of commands or process that can be automated and repeated on a bunch of images / files / stacks
File types
More info here: https://imagej.nih.gov/ij/docs/guide/146-7.html
.tif – high dynamic range image data. Always use for scientific and computing purposes
.gif – animated, low dynamic range images. Great for making animations in powerpoint.
.jpeg – compressed, very low dynamic range image data. Only use for visualization, never for computing or quantifying.
.bmp – somewhere between a tif and a jpeg. Generally not useful for scientific applications, but fine for visualization.
.nd2 – Nikons proprietary image file format. Its basically a tiff but with a bunch of handy metadata pertaining to image conditions.
