Distributed and Networked Analysis of Volumetric Image Data (DINAVID)

DINAVID is a client-server-based system that provides tools to process and render microscopy images in 3D.

DINAVID provides traditional image processing tools such as filtering, thresholding, background subtraction, and morphological operations. These operations are applied to an image volume before performing segmentation using watershed or deep-learning-based methods. DINAVID provides a deep-learning-based 3D segmentation method named DeepSynth. DeepSynth segments fluorescent microscopy images of cell nuclei using a selectable pretrained model based on the similarity between users' data and the data used to train DeepSynth.

DINAVID also provides basic cell gating tools based on the idea of flow cytometry. Users can interact with a scatterplot of measurements of the segmented results that highlight cells that correspond to the region of interest selected via a rectangular gate. In the future, we plan to implement gates of arbitrary shapes.

DINAVID also provides a three-dimensional rendering tool of microscopy images, in addition to two-dimensional rendering on a focal plane by focal plane basis. This rendering system provides the ability to perform color correction by allowing the adjustment of brightness, gamma, and offset in the RGB color system. Currently, only composite TIFF images of up to 4 channels are accepted. DINAVID will be able to render up to 7-channel composite images in the future. DINAVID will also provide the ability to perform channel attribution in the future.