Zegami is a visual data exploration tool for viewing large collections of images and other visual data. By presenting vast quantities of images within a single field of view, Zegami allows users to quickly see correlations, outliers and relationships by leveraging the innate pattern recognition capabilities of the human brain. This combined with image processing and advanced machine learning techniques like Deep Learning presents a formidable combination of the best aspects of man and machine. In addition, by including associated metadata (structured and unstructured) users can search, sort and filter the images to find exactly what is required.
In the above example researchers at the Weatherall Institute of Molecular Medicine, University of Oxford were able to quickly observe the effects of various concentrations of a particular drug against a control group.
Zegami has been built as a platform to allow it to integrate with a wide variety of data and image sources. Out of the box flat files and most relational data bases are supported, however other data sources can also be supported by leveraging Zegami’s API’s.This ability to connect to a variety of data sources and generate collections which can be viewer within a browser makes Zegami a natural platform for publishing data. Organisations that have a requirement for open data can use Zegami to .
The Australian Plant Phenomics Facility (http://www.plantphenomics.org.au/) are a world class, high-throughput phenotypying organisation which is used by both commercial and non-commercial researchers around the world to conduct research in plant sciences. Collaborators include the University of Nebraska and the Commonwealth Science and Industrial Research Organisation.
Zegami is their platform of choice to publish the tens of thousands of images they produce on a monthly basis. https://zegami.plantphenomics.org.au. Not only has this provided them a way to explore their data and make new connections which were previously time consuming, but they are able to directly link to Zegami when publishing their results.
Zegami is a new and innovative way to visualise large collections of images in a way that is currently not possible using traditional tools. This is reinforced with strong analytical functions which allow users to quickly identify correlations and outliers.
Zegami features many advanced data visualisation and analytical tools to allow users to identify patterns within the data. In addition, Zegami’s plugin API allows developers to create their own visualisations.
Zegami is continuing to develop and integrate with cutting edge technologies, specifically in Image Processing techniques and Machine Learning whether it be supervised or unsupervised.
For supervised learning, Zegami provides powerful annotation capabilities to allow domain experts to rapidly generate ground truth datasets. These datasets can then be fed into Deep Learning systems for processing, the results of which can be further analysed in Zegami. Such analysis can be used to automate the detection of various regions of interest within an image, this can also be combined with textural analysis (such as sentiment) to provide a holistic set of metadata.
Similarity search allows users to find images that look similar to a selected image purely based on visual characteristics
Combined with a strong plugin architecture, software developers can enhance Zegami with custom features, interfaces and machine learning systems.
See how Zegami works