Lumo Imaging

LumoScannner for Researchers

LumoScanner for Researchers

Many researchers around the world are feverishly working on the next generation of AI systems that can be used for better diagnosis of skin diseases/conditions, including skin cancers.
Until now, these researchers were not able to experiment with or deploy their technologies in a realistic setting. They had to either rely on a dermatologist to capture dermoscopic images (which in itself preclude any conclusion regarding the efficacy of the algorithms in the wild) or use a Total Body Scanner and make do with its suboptimal clinical image resolution for their AI analysis systems.
The lack of prior dermoscopic resolution images posed challenges in the development of temporal analysis approaches that required exact size or other feature values from the past state of the lesions. We have therefore developed a special version of LumoScanner for researchers that we call LumoScanner for Researchers.

The LumoScanner Injury Documentation System is an advanced, non-contact imaging solution designed for precise wound assessment, monitoring, and documentation.

Since the LumoScanner for Researchers can be used with an IRB, every version of every Lumo developed software, regardless of its FDA registration/clearance, can be made available to the researchers. The researcher version also allows the researchers to work with sensitive medical imaging data locally without the complexities of implementing an IT-intensive cloud implementation. 

LumoScanner for Researchers systems provides the following functionalities:

  • LumoScanner Total Body Photography (TBP) System.
  • LumoCrunch image processing pipeline that, amongst other functions, finds every lesion on the subject’s body and extracts real-sized dermoscopic-like images of every lesion on the body.
  • LumoDoc, a TBP DICOM viewer for local as well as remote patients’ scan examinations.
  • LumoTrack automated mole-mapping clinical decision support software compares every lesion from a newer scan with the same lesion from a previous scan, providing the dermatologists with information regarding the amount and kinds of transformation for every lesion on the patient’s body.

LumoScanner image collection system

As with the upcoming LumoScanner for the Clinicians, the Researchers version is a recumbent system with six 61MP cameras, a gantry, depth cameras, and a lighting system. However, the specific scanner hardware for the researchers provides the researcher with added capabilities pre-purchase special requests and ease of post purchase modifications.

For example, the research system comes with White illumination sub-system however, the researchers will be able to order it with a lighting subsystem that is capable of White, Woods Lamp, and 405nm illuminations. The Researcher system can scan each subject twice (front and back) or four times (anterior, posterior, and side views).
The LumoScanner for Researcher is also capable of supporting a large set of resolutions. Regardless of which resolution setting is selected for the study, the LumoScanner automatically computes the gantry position, camera pan angle, and focus distance for each image so that the entirety of images in aggregate provides the optimum resolution for the specific body shape of the subject.
Please see the table below for some of the possible resolution settings:

LumoCrunch image processing software

LumoCrunch cloud software uses the images from the scanner as input, processes them and outputs a DICOM file for future analysis and/or visualization. LumoCrunch.

  • Renders the large number 2D images from the scanner as one 3D rendering.
  • Extracts the raw lesion information that is useful for clinicians (e.g. crops, locations, and sizes of pigmented skin lesions)
  • Creates a TBP DICOM file that contains the 2D images, 3D rendering and lesion data.

Incidentally, Lumo Imaging has developed and published a proposal for TBP DICOM and is currently serving as a voting member of the working group that is tasked with the development of the TBP DICOM standard. Please note the crops from 2D images are orthorectified before storage in DICOM files. Orthorectification is unique to Lumo while the rest of the processing steps performed as part of the LumoCrunch pipeline are quite standard, can be readily found in literature and have been used by other vendors of TBP Systems (e.g., Fotofinder and Vectra 360).

Orthorectification is a well-known and standard step for processing satellite imaging information (akin to our senses seeing an oval-shaped tabletop and our brain interpreting the information as a circular tabletop). The resulting processed data will be stored as a DICOM, most probably in the provider’s cloud space.

LumoDoc scan visualization software

LumoDoc is full-featured cloud-based (Zero footprint) TBP DICOM viewer that can fetch DICOM files from the DICOM store using Google Health API. As mentioned before, the current unpublished TBP DICOM format contains 2D images, 3D rendering as well as crops of lesions from the 2D images.

LumoDoc can visualize all three kinds of information, allow the clinician to examine the scan from anywhere and allow the clinician to annotate the scans.

Additionally, LumoDoc can side-by-side visualize the information from two TBP DICOM files (presumably from two different points of time but from the same individual). For example the clinician can side-by-side compare two 3D renderings, two 2D images or 2 lesion crops in order to view the subtle changes in the lesions over time.

LumoDoc includes the following features:

  • A Virtual Dermatoscope that can provide highly magnified images of the lesions.
  • A Machine learning option that allows the clinician to opt for CycleGAN style transfer technology in order to render the magnified images in dermatoscopic style (as opposed to clinical style).
  • Automatic and continuous computation and display of the scale marks within the Virtual Dermatoscope.
  • Absolute and Ugly Duckling classifications are presented as heatmaps to attract the attention of the Dermatologist to significant lesions.

Ready to Experience the Future of Scanning?