Image Analysis Training Resources

Tool installation

Overview

Teaching: min
Exercises: min
Questions
Objectives

  • Install the software that is required to execute the activities in this training material

Installation instructions

To visualise the installation instructions, please select the required platform in the activity list.

Key Points

Digital image basics

Overview

Teaching: min
Exercises: min
Questions
Objectives

  • Understand that a digital image is typically stored as an N-dimensional array.

  • Learn that the image array elements are called pixels (2D) or voxels (3D).

  • Examine the values and locations of pixels/voxels in an image.

Digital image dimensions

There are several ways to describe the size of a digital image. For example, the following sentences describe the same image.

Note that “images” in bioimaging can also have more than two dimensions and one typically specifies how to map those dimensions to the physical space (x,y,z, and t). For example, if you acquire a 2-D movie with 100 time points and each movie frame consisting of 256 x 256 pixels it is quite common to represent this as a 3-D array with a shape of ( 256, 256, 100 ) accompanied with metadata such as ( (“x”, 100 nm), (“y”, 100 nm), (“t”, 1 s) ); check out the module on spatial calibration for more details on this.

Key Points

Lookup tables

Overview

Teaching: min
Exercises: min
Questions
Objectives

  • Understand how the numerical values in an image are transformed into colourful images.

  • Understand what a lookup table (LUT) is and how to adjust it.

  • Appreciate that choosing the correct LUT is a very serious responsibility when preparing images for a talk or publication.

Lookup tables do the mapping from a numeric pixel value to a color. This is the main mechanism how we visualise microscopy image data. In case of doubt, it is always a good idea to show the mapping as an inset in the image (or next to the image).

Single color lookup tables

Single color lookup tables are typically configured by chosing one color such as, e.g., grey or green, and choosing a min and max value that determine the brightness of this color depending on the value of the respective pixel in the following way:

brightness( value ) = ( value - min ) / ( max - min )

In this formula, 1 corresponds to the maximal brightness and 0 corresponds to the minimal brightness that, e.g., your computer monitor can produce.

Depending on the values of value, min and max it can be that the formula yields values that are less than 0 or larger than 1. This is handled by assinging a brightness of 0 even if the formula yields values < 0 and assigning a brightness of 1 even if the formula yields values larger than 1. In such cases one speaks of “clipping”, because one looses (“clips”) information about the pixel value (see below for an example).

Clipping example

min = 20, max = 100, v1 = 100, v2 = 200

brightness( v1 ) = ( 100 - 20 ) / ( 100 - 20 ) = 1

brightness( v2 ) = ( 200 - 20 ) / ( 100 - 20 ) = 2.25

Both pixel values will be painted with the same brightness as a brightness larger than 1 is not possible (see above).

Multi color lookup tables

As the name suggestes multi color lookup tables map pixel gray values to different colors.

For example:

0 -> black

1 -> green

2 -> blue

3 -> ...

Typical use cases for multi color LUTs are images of a high dynamic range (large differences in gray values) and label mask images (where the pixel values encode object IDs).

Sometimes, also multi color LUTs can be configured in terms of a min and max value. The reason is that multi colors LUTs only have a limited amount of colors, e.g. 256 different colors. For instance, if you have an image that contains a pixel with a value of 300 it is not immediately obvious which color it should get; the min and max settings allow you to configure how to map your larger value range into a limited amount of colors.

Key Points

  • A LUT has configurable contrast limits that determine the pixel value range that is rendered linearly.

  • LUT settings must be responsibly chosen to convey the intended scientific message and not to hide relevant information.

  • A gray scale LUT is usually preferable over a colour LUT, especially blue and red are not well visible for many people.

  • For high dynamic range images multi-color LUTs may be useful to visualise a wider range of pixel values.

Multichannel images

Overview

Teaching: min
Exercises: min
Questions
Objectives

  • Understand/visualize different image channels.

Key Points

Spatial calibration

Overview

Teaching: min
Exercises: min
Questions
Objectives

  • Understand that a pixel index is related to a physical coordinate.

  • Understand that a spatial calibration allows for physical size measurements.

Isotropy

One speaks of isotropic sampling if the pixels have the same extent in all dimensions (2D or 3D).

While microscopy images typically are isotropic in 2D they are typically anisotropic in 3D with coarser sampling in the z-direction.

It is very convenient for image analysis if pixels are isotropic, thus one sometimes resamples the image during image analysis such that they become isotropic.

Key Points

N-dimensional images

Overview

Teaching: min
Exercises: min
Questions
Objectives

  • Explore and view the different dimensions image data can have.

Key Points

Data types

Overview

Teaching: min
Exercises: min
Questions
Objectives

  • Understand that images have a data type which limits the values that the pixels in the image can have.

  • Understand common data types such as 8-bit and 16-bit unsigned integer.

Image data types

The pixels in an image have a certain data type. The data type limits the values that pixels can take.

For example, unsigned N-bit integer images can represent values from 0 to 2^N -1, e.g.

Intensity clipping (saturation)

If the value of a pixel in an N-bit unsigned integer image is equal to either 0 or 2^N - 1, you cannot know for sure whether you lost information at some point during the image acquisition or image storage. For example, if there is a pixel with the value 255 in an unsigned integer 8-bit image, it may be that the actual intensity “was higher”, e.g. would have corresponded to a gray value of 302. One speaks of “saturation” or “intensity clipping” in such cases. It is important to realise that there can be also clipping at the lower end of the range (some microscopes have an unfortunate “offset” slider that can be set to negative values, which can cause this).

Key Points

Image file formats

Overview

Teaching: min
Exercises: min
Questions
Objectives

  • Open and save various image files formats

  • Understand the difference between image voxel data and metadata

  • Understand that converting between image file formats likely leads to loss of information

Key Points

Quantitative image inspection and presentation

Overview

Teaching: min
Exercises: min
Questions
Objectives

  • Quantitatively inspect and present fluorescent microscopy images.

Key Points

Volume slicing

Overview

Teaching: min
Exercises: min
Questions
Objectives

  • Create slice views of volumetric image data

  • Master different ways of dealing with anisotropic voxels

The word ‘slice’ is often used in different ways. The different ‘layers’ in the z-dimension are referred to as z-slices. Slicing (or subsetting) image data means that part of the image data is selected and ‘sliced out’ to form a new image. This can include selecting one or more dimensions, or just part of a dimension, for example selecting slice 6-12 of the Z-dimension. You can also rotate the data in one of the spatial dimensions and resample the data set to see that data from a different angle, which is sometimes referred to as ‘reslicing’.

Key Points

Projections

Overview

Teaching: min
Exercises: min
Questions
Objectives

  • Project multi-dimensional image data into lower dimensions

  • Understand the differences between projection modes such as max, sum, and mean

Key Points

Volume rendering

Overview

Teaching: min
Exercises: min
Questions
Objectives

  • Understand the concepts and some methods of 3-D rendering.

  • Appreciate that 3-D rendering can be challenging for some data.

  • Perform basic volume rendering using a software tool.

Volume rendering software

Software Multi-Channel Time-lapse Max-Projection Volume Iso-Surface … … … …
Blender                  
Drishti                  
ImageJ 3Dscript                  
ImageJ 3D viewer N N N Y Y        
ImageJ ClearVolume (Upate Site) Y Y Y N N        
ImageJ Volume Viewer N N Y Y N        
Napari                  

Key Points

Segmentation

Overview

Teaching: min
Exercises: min
Questions
Objectives

  • Understand the difference between instance and semantic segmentation

Key Points

Thresholding

Overview

Teaching: min
Exercises: min
Questions
Objectives

  • Describe the relationship between an intensity image and a derived binary image

  • Apply a threshold to segment an image into foreground and background regions

A common algorithm for binarization is thresholding. A threshold value t is chosen, either manually or automatically, and all pixels with intensities below t are set to 0, whereas pixels with intensities >= t are set to the value for the foreground. Depending on the software the foreground value can be different (e.g. 1 in MATLAB or 255 in ImageJ). At any pixel (x,y):

p_im(x,y) < t -> p_bin(x,y) = 0

p_im(x,y) >= t -> p_bin(x,y) = 1

where, p_im and p_bin are the intensity and binary images respectively.

It is also possible to define an interval of threshold values, i.e. a lower and upper threshold value. Pixels with intensity values within this interval belong to the foreground and vice versa.

Key Points

Automatic thresholding (histogram-based)

Overview

Teaching: min
Exercises: min
Questions
Objectives

  • Understand how an image histogram can be used to derive a threshold

  • Apply automatic threshold to distinguish foreground and background pixels

Key points

Key Points

Connected component labeling

Overview

Teaching: min
Exercises: min
Questions
Objectives

  • Understand how objects in images are represented as a label mask image.

  • Apply connected component labeling to a binary image to create a label mask image.

Typically, one first categorise an image into background and foreground regions, which can be represented as a binary image. Such clusters in the segmented image are called connected components. The relation between two or more pixels is described by its connectivity. The next step is a connected components labeling, where spatially connected regions of foreground pixels are assigned (labeled) as being part of one region (object).

In an image, pixels are ordered in a squared configuration.

For performing a connected component analysis, it is important to define which pixels are considered direct neighbors of a pixel. This is called connectivity and defines which pixels are considered connected to each other.

Essentially the choice is whether or not to include diagonal connections.

Or, in other words, how many orthogonal jumps to you need to make to reach a neighboring pixel; this is 1 or an orthogonal neighbor and 2 for a diagonal neighbor.

This leads to the following equivalent nomenclatures:

Key Points

Object shape measurements

Overview

Teaching: min
Exercises: min
Questions
Objectives

  • Understand shape measurements and their limitations

  • Perform shape measurements on objects

Key Points

Nuclei segmentation and shape measurement

Overview

Teaching: min
Exercises: min
Questions
Objectives

  • Create a basic image analysis workflow.

  • Understand that bioimage analysis workflows consist of a sequence of image analysis components.

  • Segment nuclei in a 2D image and measure their shapes and understand the components (concepts and methods) that are needed to accomplish this task.

  • Draw a biophysically meaningful conclusion from applying an image analysis workflow to a set of images.

Key Points

Fluorescence microscopy image formation

Overview

Teaching: min
Exercises: min
Questions
Objectives

  • Understand how the intensities in a digital image that was acquired with a fluorescence microscope are formed

  • Understand how this image formation process has critical influence on the interpretation of intensity measurements

  • Understand that the geometry of your biological specimen can have a large influence on the measured intensities

Key Points

Object intensity measurements

Overview

Teaching: min
Exercises: ImageJ GUImeasure_intensities/exercises/measure_intensities_imagejgui.md min
Questions
Objectives

  • Understand the correct biophysical interpretation of the most common object intensity measurements

  • Perform object intensity measurements, including background subtraction

Nomenclature

Formula

mean_corr = mean - bg
sum_corr = mean_corr * num_pixels = ( mean - bg ) * num_pixels = sum - ( bg * num_pixels )

Biophysical interpretation

Key points

Key Points

Global background correction

Overview

Teaching: min
Exercises: ImageJ Macro & GUIglobal_background_correction/exercises/global_background_correction.md min
Questions
Objectives

  • Measure the background in an image

  • Apply image math to subtract a background intensity value from all pixels and understand that the output image should have a floating point data type

Key Points

Neighborhood filters

Overview

Teaching: min
Exercises: min
Questions
Objectives

  • Understand the basic principle of a neighborhood filter.

  • Apply basic neighborhood filters to an image.

Neighborhood filters comprise two ingredients: a definition of the pixel neighborhood (size and shape) and a mathematical recipe what to compute on this neighborhood. The result of this computation will be used to replace the value of the central pixel in the neighborhood. This procedure can be applied to several (all) pixels of an image to obtain a filtered image. The animation shows a square neighborhood (3x3) applied to the inner pixels of the image.

There are tons of different neighborhood filters, and you can also invent one!

The neighborhoods

The neighborhood of a pixel is also called a structuring element (SE) and can have various sizes and shapes. Here, we use one of the simplest and most widely used neighborhoods, namely a circular neighborhood, which is defined by a certain radius. We will explore other shapes of SEs in more detail in a dedicated module.

Padding

Since the filtering operation takes into account all the directions of extent of SE, border pixels would be affected in a different way and one has to decide that which values they should assume. Padding is the operation of adding an additional layer around your data to get more accurate values after filtering process. It also gives you option to retain same dimensions for your data after filtering. Common padding methods are using zeros or to mirror/replicate the border pixel values.

The math

There are many ways how to cleverly compute on a pixel neighborhood. For example, one class of computations is called convolutional filters, another is called rank filters. Here, we focus on the relatively simple mean and variance filters.

Best practice

As usual, everything depends one the scientific question, but maybe one could say to use a filter that changes the image as little as possible.

Key Points

Median filter

Overview

Teaching: min
Exercises: min
Questions
Objectives

  • Understand in detail what happens when applying a median filter to an image

Properties of median filter

The median filter is based on ranking the pixels in the neighbourhood

In general, for any neighbourhood filter, if the spatial extend of the neighbourhood is significantly (maybe three-fold) smaller than the smallest spatial length scale that you care about, you are on the safe side.

However, in biology, microscopy images are often containing relevant information down to the level of a single pixel. Thus, you typically have to deal with the fact that filtering may alter your image in a significant way. To judge whether this may affect your scientific conclusions you therefore should study the effect of filters in some detail.

Although a median filter typically is applied to a noisy gray-scale image, understanding its properties is easier when looking at a binary image.

From inspecting the effect of the median filter on above test image, one could say that a median filter

Key Points

2D noisy object segmentation and filtering

Overview

Teaching: min
Exercises: min
Questions
Objectives

  • Create an image analysis workflow comprising image denoising and object filtering.

Key Points

Morphological filters

Overview

Teaching: min
Exercises: min
Questions
Objectives

  • Understand how to design morphological filters using rank filters

  • Execute morphological filters on binary or label images and understand the output

Rank filters

In the region defined by the structuring element, pixel elements are ranked/sorted according to their values. The pixel in the filtered image is replaced with the corresponding sorted pixel (smallest = min, greatest = max, median ). See also Median filter. Morphological filters corresponds to one or several rank filters applied to an image.

Morphological filters on binary images

A typical application of these filters is to refine segmentation results. A max-filter is called dilation whereas a min-filter is called erosion. Often rank filters are applied in a sequence. We refer to a closing operation as a max-filter followed by a min-filter of the same size. An opening operation is the inverse, a min-filter followed by a max-filter.

Opening operations will:

Closing operations:

Image subtraction using eroded/dilated images allows to identify the boundary of objects and is referred to morphological gradients:

Fill holes operation is a slightly more complex morphological operation. It is used to identify background pixels surrounded by foreground pixels and change their value to foreground. Algorithmically there are several ways to achieve this.

Morphological filters on label images

Morphological filters work also on label images. If the objects are not touching this will achieve the expected result for each label. However, when objects touch each other, operations such as dilations can lead to unwanted results.

Morphological filters on grey level images

Min and max operations can be applied to grey level images. Applications are for example contrast enhancement, edge detection, feature description, or pre-processing for segmentation.

Key Points

Local background correction

Overview

Teaching: min
Exercises: min
Questions
Objectives

  • Understand how to use image filters for creating a local background image

  • Use the generated local background image to compute a foreground image

There exist multiple methods on how to compute a background image. Which methods and parameters work best depends on the specific input image and the size of the object of interest.

Common methods are:

Some of the methods may be sensistive to noise. Therefore, it can be convenient to smooth the image, e.g. with a mean or gaussian filtering, prior computing the background image.

Key Points

Object filtering

Overview

Teaching: min
Exercises: ImageJ GUIfilter_objects/exercises/filter_objects_imagejgui.md min
Questions
Objectives

  • Remove objects from a label mask image.

Key Points

Distance transform

Overview

Teaching: min
Exercises: Distance to center, ImageJ GUIdistance_transform/exercises/distance_transform_geodesic_imagejgui.mdGlial thickness, ImageJ GUIdistance_transform/exercises/distance_transform_skeldist_imagejgui.md min
Questions
Objectives

  • Understand how to use distance transform to quantify morphology of objects

  • Understand how to use distance transform to quantify distance between objects

  • Understand approximate nature of distance transform

Chamfer distance (modified from MorpholibJ Manual)

Several methods (metrics) exist for computing distance maps. The MorphoLibJ library implements distance transforms based on chamfer distances. Chamfer distances approximate Euclidean distances with integer weights, and are simpler to compute than exact Euclidean distance (Borgefors, 1984, 1986). As chamfer weights are only an approximation of the real Euclidean distance, some differences are expected compared to the actual Euclidean distance map.

Several choices for chamfer weights are illustrated in above Figure (showing the not normalized distance).

To remove the scaling effect due to weights > 1, it is necessary to perform a normalization step. In MorphoLibJ this is performed by dividing the resulting image by the first weight (option Normalize weights).

Key Points

Watershed

Overview

Teaching: min
Exercises: ImageJ Macro: MorpholibJ shape watershedwatershed/exercises/morpholibj_shape_watershed_exercise.ijmImageJ Macro: MorpholibJ seeded watershedwatershed/exercises/morpholibj_seeded_watershed_exercise.ijm min
Questions
Objectives

  • Understand the concept of watersheds in image analysis.

  • Understand that a watershed algorithm is often applied to a distance map to split objects by their shape.

  • Be able to run a watershed algorithm in an image analysis platform.

Key Points

  • A watershed transform can separate touching objects if there are intensity valleys (or ridges) between touching objects. In case of intensity ridges the image needs to be inverted before being subjected to the watershed transform.

  • To separate object by their shape, use a distance transform on the binary image and inject this into the watershed transform. It is often good to smooth the distance transform to remove spurious minima, which could serve as wrong seed points and thus lead to an over-segmentation.

Nuclei and cells segmentation

Overview

Teaching: min
Exercises: min
Questions
Objectives

  • Segment cells and nuclei, using nuclei as seeds for watershed segmentation of the cells.

Key Points

Skeletonization

Overview

Teaching: min
Exercises: ImageJ GUIskeletonization/exercises/skeletonization_imagejgui.mdmarkdownImageJ Macroskeletonization/exercises/skeletonization_imagejmacro.mdjavaImageJ Jythonskeletonization/exercises/skeletonization_imagej-Jython.mdpython min
Questions
Objectives

  • Apply a skeletonization algorithm to a binary image to view its internal skeleton

  • Count the number of branches and branch lengths to obtain morphological information from the image

Key Points

Similarity transformations

Overview

Teaching: min
Exercises: min
Questions
Objectives

  • Understand how similarity transformations alter an image

  • Understand that similarity transforms may create new pixels, which must be created using a carefully chosen interpolation mode

  • Similarity transform an image

Key Points

OME-Zarr

Overview

Teaching: min
Exercises: min
Questions
Objectives

  • Understand the OME-Zarr image file format

  • Render cloud (S3 object store) hosted OME-Zarr image data

  • Access the pixel values of cloud hosted OME-Zarr image data

  • Apply basic image processing on cloud hosted OME-Zarr image data

Key Points

Running a script

Overview

Teaching: min
Exercises: ImageJ Macro in Fijiscript_run/exercises/script_run_fiji_imagej_macro.md min
Questions
Objectives

  • Understand that a script is a single text file that is written in a specific scripting language

  • Understand the basic building blocks of a script, i.e. what happens in each line

  • Run a bioimage analysis script

  • Modify a bioimage analysis script

Programming script content

A programming script is a text file where each line is code that can be executed by the platform (the compiler) in which you are running the script. There are different types of content that a line can represent. Sometimes one line can even contain multiple of such contents. In the following sections some of the very common types of content are very briefly discussed (check out the follow-up modules for much more details).

Comments

It is good practice to add some human readable comments to explain what the code is doing. To tell the compiler that a part of a script is a comment, one prepends the comment section special symbol, such as // or #.

Examples:

Import statements

In some cases one needs to tell the executing environment which libraries are needed to run the code. This is done via so-called import statements.

Examples:

Functions and parameter

Functions are the heart of a program, they do stuff, depending on the paramteres that you give.

Examples:

Variables

Very often you want to store the results of some computation. In most languages this is achieved by the = sign operator, where you assign the right of the = sign to the varaible on the left.

Examples:

Key Points

Coding with LLMs

Overview

Teaching: min
Exercises: min
Questions
Objectives

  • Use a large language model (LLM) to create bioimage analysis code

  • Use a LLM to understand bioimage analysis code

Key Points

Recording a script

Overview

Teaching: min
Exercises: ImageJscript_record/exercises/script_record_imagejgui.md min
Questions
Objectives

  • Record graphical user interface (GUI) actions into a script

Key Points

Variables

Overview

Teaching: min
Exercises: ImageJ Macro, difference of gaussiansscript_variables/exercises/script_variables_DoG_imagejmacro.mdImageJ Macro, fix itscript_variables/exercises/script_variables_fixit_imagejmacro.md min
Questions
Objectives

  • Understand difference between variable name, value, type, and storage

  • How to use variables in functions

Key Points

Working with strings

Overview

Teaching: min
Exercises: ImageJ Macro: Concatenate variablesstring_concat/exercises/string_concat_imagejmacro.mdImageJ Macro: Create function argumentsstring_concat/exercises/string_concat_imagejmacro2.mdImageJ Macro: Create pathsstring_concat/exercises/string_concat_imagejmacro3.md min
Questions
Objectives

  • Construct complex strings, e.g. to produce log messages and create file paths

  • Master the backward slash .

Creating paths

A frequent operation in bioimage analysis is to create paths to images by concatenating a folder and file name to a full path. Please note that when concatenating a folder and a file name into a full path, you might need to add a so-called file separator between the folder and the file name. This is a character that separates directory names within a path to a particular location on your computer. Different operating systems use different file separators: on Linux and MacOS, this is /, while Windows uses \. To make it worse, when you store a directory you are typically never sure whether the contained string ends on / or \ or does not have the separator in the end, e.g. C:\Users\Data, in which case you have to add it when concatenating a file name). To make it even worse, in some programming langauges the \ character have a special meaning within strings and is thus not simply interpreted as a character and to actually get a backslash you may have to write \\.

If you want to have some “fun” you can read those discussions:

As all of this can quickly become a huge mess, fortunately, scripting languages typically offer special functions to help you write code to create file paths that will run on all operating systems.

String concatenation

String concatenation is the operation of joining multiple substrings.

For example concatenating “Hello “ and “world!” would result into “Hello world!”.

Key Points

Output saving

Overview

Teaching: min
Exercises: ImageJ Macrooutput_saving/exercises/output_saving_imagej-macro.mdImageJ Jythonoutput_saving/exercises/output_saving_imagej-jython.md min
Questions
Objectives

  • Save measurements as a table

  • Save ROIs

  • Save output label mask

Key Points

Batch processing

Overview

Teaching: min
Exercises: ImageJ Macro Scijavabatch_processing/exercises/imagejmacro.md min
Questions
Objectives

  • Automatically process a number of images

Key Points

Handling script parameters

Overview

Teaching: min
Exercises: Exposing script parameters: ImageJ Macrofetching_user_input/exercises/fetch_user_input_imagejmacro.mdmarkdown min
Questions
Objectives

  • Organise script parameters in the code such that can be easliy adapted

  • Create dialog boxes for fetching script parameters

Key Points

Commenting

Overview

Teaching: min
Exercises: min
Questions
Objectives

  • Understand the concept and purpose of commenting.

  • Comment properly what certain section of code.

Key Points

Setting up a scripting environment

Overview

Teaching: min
Exercises: min
Questions
Objectives

  • Set up a scripting environment for your platform

Key Points

Functions

Overview

Teaching: min
Exercises: Exercise1 - ImageJ Macroscript_functions/exercises/functions_imagejmacro.mdExercise2 - ImageJ Macroscript_functions/exercises/functions_imagejmacro2.md min
Questions
Objectives

  • Avoid code duplication using functions.

  • Understand the skeleton of a function.

  • Make the script more efficient.

Key Points

Loops

Overview

Teaching: min
Exercises: ImageJ Macro, Multiple erosionscript_for_loop/exercises/script_for_loop_erodeband.md min
Questions
Objectives

  • Use for loops to repeat operations multiple times

  • Running a script for multiple files

For loop

A for loop occurs by iterating over a loop variable defined in a loop header. You use for loops when you know the number of iterations to execute.

While loop

While loop does not have a fixed number of iterations. Typically the header contains a condition that is computed within the body of the loop. TODO.

Key Points

Correlative image rendering

Overview

Teaching: min
Exercises: min
Questions
Objectives

  • Understand how heterogeneous image data can be mapped from voxel space into a global coordinate system.

  • Understand how an image viewer can render a plane from a global coordinate system.

Key Points

Deep learning instance segmentation

Overview

Teaching: min
Exercises: CellPose GUIdeep_learning_run_segmentation/exercises/deep_learning_cellpose_gui.mdmarkdown min
Questions
Objectives

  • Run a deep learning model for instance segmentation

  • Visually assess the quality of the segmentation

  • Appreciate that, in order to work well, the model has to match the input data

  • Appreciate that even deep learning models may have parameters that need to be tuned

  • Appreciate that subtle differences in the image data may cause the model to fail

Key Points

Manual segmentation

Overview

Teaching: min
Exercises: min
Questions
Objectives

  • Manually segment parts of a 2-D (3-D) image.

Manual segmentation considerations

How to deal with objects that are not fully in the image?

Should objects be separated by background pixels?

Key Points

Segment Golgi objects per cell

Overview

Teaching: min
Exercises: min
Questions
Objectives

  • Segment intracellular objects and assign them to their parent cell

Key Points

Big image data file formats

Overview

Teaching: min
Exercises: min
Questions
Objectives

  • Understand the concepts of lazy-loading, chunking and scale pyramids

  • Know a concrete file format that implements chunking and scale pyramids

Key Points