OpenSegSPIM is a free Software under BSD License and is available for download as executable for Windows or Mac or any computer with a Matlab license (Matlab R2103a +).
It has been tested on a 8 core PC with 32 Gb Ram , and requires ~500Mb free space for executable version , or ~5Mb for just source code.
It has been tested on a 8 core PC with 32 Gb Ram , and requires ~500Mb free space for executable version , or ~5Mb for just source code.
Here is a step by step user guide video showing how to segment a 3D nuclei cluster sample:
The pipeline consists of the following steps:
1. Load image:
Allowing the user to load the stack of interest and enter, if needed, the appropriate calibration of the image.
2. Image Enhancement:
Three steps are included in this section: A background noise removal step, intensity adjustment and intensity smoothing.
The smoothing depends only on the nuclei diameter. The main objective of the smooth image is to improve the subsequent detection of seeds inside each nucleus.
3. Foreground extraction:
OpenSegSPIM applies an automatic thresholding (Otsu thresholding) to distinguish the foreground (i.e. intensity signal that correspond to object to be detected) and background (remaining noise and empty background). The Otsu thresholding parameter automatically adjusted is multiplied by a coefficient adjustable by the user.
4. Detection:
Once the foreground is extracted after Gaussian smoothing, the software will detect each different object in foreground. Two different approaches are available:
-An intensity based seeds detection where local maxima of intensity are detected. To be used when the signal intensity is good and homogeneous within the objects to be detected (the nucleus)
-A shape based seeds detection using the distance function of the object. A pseudo 3D preview is available to quickly compare the 2 different approaches on a reduced size stack.
5. Segmentation:
Watershed is applied either on the intensity image or the shape image depending on which approach was used for the seeds detection.
6. Analysis:
Once all detected objects have been segmented the user can measure different parameters such as volume, sphericity, nearest neighbour distance and total intensity for each object.
For each step OpenSegSPIM gives a direct visualisation of the result and allows the user to repeat each step until the result is satisfying. The final result is saved as multiple tiff stacks and statistical measurement are saved as .txt file for further use.
1. Load image:
Allowing the user to load the stack of interest and enter, if needed, the appropriate calibration of the image.
2. Image Enhancement:
Three steps are included in this section: A background noise removal step, intensity adjustment and intensity smoothing.
The smoothing depends only on the nuclei diameter. The main objective of the smooth image is to improve the subsequent detection of seeds inside each nucleus.
3. Foreground extraction:
OpenSegSPIM applies an automatic thresholding (Otsu thresholding) to distinguish the foreground (i.e. intensity signal that correspond to object to be detected) and background (remaining noise and empty background). The Otsu thresholding parameter automatically adjusted is multiplied by a coefficient adjustable by the user.
4. Detection:
Once the foreground is extracted after Gaussian smoothing, the software will detect each different object in foreground. Two different approaches are available:
-An intensity based seeds detection where local maxima of intensity are detected. To be used when the signal intensity is good and homogeneous within the objects to be detected (the nucleus)
-A shape based seeds detection using the distance function of the object. A pseudo 3D preview is available to quickly compare the 2 different approaches on a reduced size stack.
5. Segmentation:
Watershed is applied either on the intensity image or the shape image depending on which approach was used for the seeds detection.
6. Analysis:
Once all detected objects have been segmented the user can measure different parameters such as volume, sphericity, nearest neighbour distance and total intensity for each object.
For each step OpenSegSPIM gives a direct visualisation of the result and allows the user to repeat each step until the result is satisfying. The final result is saved as multiple tiff stacks and statistical measurement are saved as .txt file for further use.