Unlock the most cutting-edge data types and integration possibilities in this future-facing session. We’ll demo how to explore your data in X Reality and gaming engines, plus give an overview of exciting new directions for the FME development team.
6. Indoor Mapping Challenges
● Must integrate multiple sources to
produce an indoor map.
○ GeoJSON, Revit, IFC, CAD (Autodesk,
Bentley), Civil 3D, Esri Geodatabase,
databases, CityGML …
● Must transform inconsistent data.
● Must comply with specifications of
the indoor format, e.g. IMDF, HERE,
ArcGIS Indoors, IndoorGML.
○ Strict data models and explicit spatial
relationships.
● Venues constantly change, so maps
need to be updated automatically.
10. Scenario: Augmented Wayfinding
1. Create floor network dataset.
2. Use the ShortestPathFinder.
3. Write to FMEAR format.
4. Make a webpage that opens
the FME AR app.
11. Scenario: Augmented Objects
1. Create/load your floor plan.
2. Change existing objects,
e.g.color or texture.
3. Add new objects, e.g.
furniture, annotations,
action heroes.
12. Scenario: Virtual Scenes
1. Read a 3D model.
2. Use the ThreejsSceneCreator.
3. Write the scene and host it
locally or on the web.
4. Put on your headset and
explore!
17. ● Can we analyze brain scan data (CT, MRI) to
identify brain tumors?
○ Knowing a tumor’s location and type would
help reduce radiation damage to healthy tissue.
● Manual approach is prone to human error and
time consuming.
● Helpful FME capabilities:
1. Process huge volumes of images.
2. Do object detection, which would help isolate
a tumor’s location.
3. Classify data, which would help identify the
tumor subtype.
Goal: Identify Brain
Tumors
18. DICOM images in time (left) and space (right)
DICOM images in time (left) and space (center/right)
24. FME Data Express
Making it easy for anyone to run FME Workspaces as
apps on their mobile device.
● Leverage device info like location and camera.
● Control permissions with tokens.
Technology trends we’re diving into at Safe Software.
Note to presenters: The key for this presentation is to tell the audience why they should care about these emerging technologies and how FME can help them embrace it. Don’t make it a self-indulgent talk (“this is what we care about at Safe”) – rather, make it about the audience (this is a trend and this is why you should care about it).
Rich building information models are widespread. Being able to integrate and work with BIM is more important than ever.
This means you can transform and convert Revit data easily. Connect directly to the source. Before now, Revit wouldn’t let software connect directly to the source files.
As mentioned in the Quality Control talk, venues worldwide are generating indoor maps of their spaces to track/geolocate assets and help patrons navigate.
Integrate: Often multiple data sources need to be combined.
Transform: Revit vs. 2D floorplans vs. other sources that don’t match up with the indoor mapping format’s requirements.
Comply: We looked at this in the QA talk.
Automate: Many venues change constantly - airports, conference centres. Need ‘live’ updates.
Keep your indoor maps in sync by setting up your workflow to run as new data arrives (FME Server).
You can use FME to do all of this! Integrate sources, transform it to meet requirements, automatically keep indoor maps synchronized.
FME helps you meet strict requirements for all of these indoor mapping formats. Convert and validate.
Left: Victoria airport. Right: UMass Amherst.
Dozens of Airports, Campuses, and other venues are already using FME for indoor mapping.
Airports use FME to track assets on indoor maps (helps operate efficiently) and for wayfinding (helps improve passenger experience).
They use FME to read floor plans, asset databases, and other key information, transform it to meet indoor mapping requirements, and then write it to their indoor mapping format of choice. Then they can use it in popular apps like Apple Maps, in a custom app (like Vancouver airport does), on their website (like Schiphol) … The final step is to automate the workflow: venues like airports change constantly due to construction and changing foot traffic, so it’s important to generate new indoor maps regularly to reflect these changes. (On a schedule or by watching for changing source data.)
A transition to the next topic could go like this - there is not much new for FME in the data preparation process for indoor with the exception of formats and standards themselves - we always were strong in combining all kinds of data and bringing it together. What is new is how we interact with these new indoor environments, and X Reality is a good way of doing it.
“X” means Virtual, Augmented, and Mixed.
Experiencing data in an immersive way is becoming more popular. It’s easier for people to understand data if they can see it in an XR environment. E.g. help stakeholders understand a construction project by holding up a tablet and seeing the plan augmented on the construction site, rather than just showing them a CAD file. Helps reduce design coordination errors in Architecture, Engineering, and Construction.
FME has capabilities for converting your data to XR. We also have an AR app for mobile devices.
Let’s look at 3 scenarios for how you can get your data into XR.
ThreejsSceneCreator is an FME Hub transformer https://hub.safe.com/transformers/threejsscenecreator
“Put on your headset” can also mean putting your phone into a Cardboard.
Gaming engines let you explore your datasets in the most immersive way yet! It’s impossible not to find BIM data exciting!
Go to the game in demo folder and play a bit in full screen mode. Currently, Windows only.
BC Cancer, SFU NeuroTech, and Safe Software are working together to use data integration and machine learning technology to analyze brain scan data and identify potential brain tumours with more accuracy than ever before.
This saves doctors time. It reduces errors. It helps reduce tissue damage because tumors can be correctly classified and their locations identified – so the correct amount of radiation can be directed at the exact area the tumor is in.
For doctors, having to analyze massive amounts of patient data (scans + huge tables/databases of info) is an everyday problem, so we’re looking at how we can help them save time by automatically processing it and delivering the insights they need.
DICOM data can come as a series of images - either in time or space.
For FME, there is no difference whether the data it processes is geospatial or not - as soon as we add a new format and can bring data in, we have a huge toolset to do with this data whatever we want.
“Space”: The skull is made of ~100 images taken about a mm apart.
Explanation:
Top right image is an animated GIF from the original DICOM data in the original direction of the scan
Middle and Bottom right. Image converted to ‘Point Cloud’ and then rotated and then sliced in a different orientation -> animated GIF
Coal mining companies also use CT (Computerized Tomography) scans on coal to identify how gas flows through it.
Given a CT scan of coal (a DICOM image), FME can analyze the raster cells and determine how the density varies across the coal. More vertical “cleats”, i.e. lines or planes in the coal that are less dense, mean the gas will flow more easily.
FME can automatically detect the presence of cleats using raster object detection. Raster expression evaluation can also identify patterns in the coal so mining companies know how gas will flow through it. This is huge.
Using FME to analyze raster scans and automatically find patterns and identify where to drill and in which direction.
there are new transformers for training FME to classify text excerpts, and subsequently for asking FME to classify text based on how you’ve trained it.
NLPTrainer
NLPClassifier
New transformers can be used to train FME to recognize objects, and subsequently you can pass FME huge volumes of images and FME will be able to tell you which images have those objects in them.
Here is an example where these transformers were used to train FME to recognize stop signs.
Anyone can try out the web service.
A short video https://youtu.be/9iDtR6kdAwg
The map of stop signs - https://s3-us-west-2.amazonaws.com/safe-scenarios/Leaflet/RasterObjectDetection/stopsigns.html
RasterObjectDetectorSampleGenerator
RasterObjectDetectorSamplePreparer
RasterObjectDetectionModelTrainer
RasterObjectDetector
From the opening slides. This is about the “workspaces as apps” point.
(In 2019, you can also read FME AR data.)
Now, creating an Android or iPhone app does not require special skills - it is as simple as creating a workspace.