Fig 1.
Task-1 includes the development of a pipeline for converting proprietary high-resolution microscope imaging modalities data into DICOM standard format. Task-2 includes the development of a 3D multidimensional high-resolution microscope image viewer. Task-3 includes the development of automatic content discovery for microscope imaging modalities, specifically CLSM and FIB-SEM images.
Fig 2.
The four-channel tibia bone [BL6-SEC5] of mice [27] (Perilipin 1 (visualized in red), Tyrosine Hydroxylase (visualized in green), DAPI (visualized in blue), calcitonin gene-related peptite (visualized in pink)).
Fig 3.
Framework for an interactive 3D viewer.
(a) Acquisition of multidimensional images and conversion to the DICOM standard; (b) Development of a web server using the Django web-framework; and (c) Implementation of visualization techniques in the client-side framework.
Fig 4.
Graphical user interface (GUI) of the IMAGE-IN multidimensional 3D visualizer for visualizing CLSM and FIB-SEM data.
Fig 5.
Flowchart depiciting the inner working mechanism of our viewer.
Step 1 describes how to host the server locally, and Step 2 explains how the viewer works, from describing the desired viewer to clicking the render button.
Fig 6.
Pipeline showing image conversion to volume rendering.
The OME-Bioformat library was used as a file reader to read microscope imaging raw files, and the Pydicom Python library was used to convert them into DICOM standard, which was then displayed in volume view using our proposed IMAGE-IN [24] viewer.
Fig 7.
(a) Example of a three-channel human islet microvasculature [42] CLSM dataset volume rendered in (b) Each channel is rendered with a different color: Channel-1 is cyan, Channel-2 is magenta, and Channel-3 is green.
Table 1.
Selected CLSM and FIB-SEM microscope images.
Fig 8.
The two-channel submandibular ganglion [45] CLSM dataset is rendered in 3D volume on the following viewers: (a) Clearvolume (the ImageJ open-source desktop viewer), (b) IMAGE-IN (the web visualizer), (c) Napari (the Python desktop visualizer), and (d) Imaris (the closed-source desktop viewer). The yellow arrow on each rendered image indicates the variation in region across four viewers. The image-in viewer reveals more nanosensor (represented by red color) on this specific spot when compared to the other three viewers. This comparison is made in the early phases.
Fig 9.
The three-channel mouse proximal colon [44] CLSM dataset is rendered in 3D volume on the following viewers: (a) Clearvolume (the ImageJ open-source desktop viewer), (b) IMAGE-IN (the web visualizer), (c) Napari (the Python desktop visualizer), and (d) Imaris (the closed-source desktop viewer). The orange arrow on each rendered image indicates the variation in region across four viewers. When compared to the other three viewers, the IMAGE-IN viewer clearly reveals more Syp (represented by the red color) and HU (represented by the blue color) on this specific spot. This comparison is made in the early phases.
Table 2.
3D viewer performance measurement on laptop and desktop devices.
Fig 10.
The three-channel human islet microvasculature [42] and four-channel tibia bone tissue [BL6-SEC5] [27] CLSM dataset is rendered in both (a, c) volume and (b, d) surface rendering views. In each view, each channel is rendered in a different color.
Fig 11.
The two-channel adipose tissue [43] and four-channel tibia bone tissue [C3H-SEC9] [27] CLSM dataset is rendered in both (a, c) volume and (b, d) surface rendering views. In each view, each channel is rendered in a different color.
Fig 12.
(a) and (b) show the volume rendering of Plasmodium falciparum [48] and Parasitophorous vacuole [49] FIB-SEM microscope imaging, respectively, whereas (c) shows the tri-planar view of Tuwongella immobilis [47], and (d) shows the multiplanar reconstruction view of Plasmodium falciparum [48] FIB-SEM microscope imaging modality.
Table 3.
3D viewer performance measurement on Ipad device.
Fig 13.
The two-channel mouse proximal colon [46] CLSM dataset is rendered into tri-planar viewer.
(a) Channel-1: (green) neuronal nitric oxide synthase (nNOS) (b) Channel-2: (red) neurofilament (NF) heavy chain.