Student: Sumit Kumar Vohra
Mentor: Dimiter Prodanov


International Neuroinformatics Coordinating Facility

Project description

Active Segmentation plugin is the redesign of existing Trainable Weka Segmentation (TWS) of ImageJ. The platform was developed in the context of GSOC 2016. Active Segmentation was developed with the main goal of providing a general purpose workbench that would allow biologists to access state-of-the-art techniques in machine learning to improve their image segmentation results. The Active Segmentation provides generic functionality and user friendly interface so that the user can include the state of the art filters and machine learning frameworks from the WEKA library:

The generic framework can be used for segmentation, classification and object detection. The platform is still under development although the one of main functionality(Segmentation) has been completed.

Developers should look at the IFilter, IDataset, IFilterManager, IFeature, ILearning interfaces inside active segmentation to get an idea how to create their own filters, learning methods to integrate them with plugin. Developers level document will be provided soon to get familiar with architecture of the plugin.

Session Screen

Session is a provision for storing project data in the session file. It save the state of the system. First image illustrates the window for Session Screen. Session screen will load the existing(Previous) session of the project. It provide you the option to load the filter, feature and learning data. The "Choose Session File" button is used for loading session file from the local file system. The whole meta-data of the project will be saved in the same directory as selected by the user. The session file must be in "JSON" format.The second image shows the window for browsing session file on the local file system:


Code implementing the loading of session file can be found in the "getMetaInfo" method from the "io" package.

	public MetaInfo getMetaInfo() {
			ObjectMapper mapper = new ObjectMapper();
			try {
				metaInfo= mapper.readValue(new File(path+Common.FILENAME), MetaInfo.class);
				originalImage= IJ.openImage(metaInfo.getPath()+Common.TRAININGIMAGE+Common.TIFFORMAT);
				return metaInfo;

			} catch (JsonGenerationException e) {
			} catch (JsonMappingException e) {
			} catch (IOException e) {

			metaInfo= new MetaInfo();
		return metaInfo;

As can be seen from the "ObjectMapper" in the code, the session file was mapped using the Jackson API, version 2.3.1.

The following code shows the format of session file for filters. FilterList consists of meta data information related to all the filters which are currently available in active segmentation. Each element in the list will give information about individual Filter. The key field in Filter is FilterName that will be unique for every filter. Filter will also consist of settings in key- value pair format.

		"FilterList": [
		{"Filter":"FilterName1","Setting1":"true","Setting2":"2","Setting3":"false", "FileList": "FilterName1.tif"},
		"FileList": "FilterName2.tif"}

Filter Framework

Filter Framework is responsible for automatic loading of filters.It automatically finds all JAR files in the path and first level of subdirectories and those have implemented "IFilter" interface.

Active Segmentation Filters Screen

Feature Screen

Feature Screen of the active segmentation is very similar to the TWS(Trainable Weka Segmentation). It will start with two classes i.e Binary Classification

Active Segmentation Features Screen

Configure Screen

Configure screen is used to configure classes. The user can add,update or delete classes. The default classes will be two.