Challenge: analyzing vehicle headlights with CP or CPA


#1

Hello everyone,
I have a difficult challenge which I hope you can help me with. I am trying to distinguish between the headlight patterns of different vehicles (see picture), and it seemed to me that these patterns kinda look like cells. Is there any way to somehow characterize the headlight patterns, and then distinguish between different vehicle types?
Many thanks in advance!
zpast


#2

Neat! Not a bad idea. Do you have images that lack wording and are individual available? It will be much easier to build a pipeline with those.


#3

Hi there,

You can use UnmixColors to extract the most contrast image of the headlights
Then invert the resulting images, for something like this
33 PM

Then identify the bright objects as usual.

Or, better, please look for computer vision solutions.

Good luck


#4

Thank you for your replies! I am a novice so would really appreciate any explanations and procedures.
Here are some pictures again, individually and without overlayed texts.

Z.

!


#5

Hi there,

I guess you already had CellProfiler installed and running?
Then please load these images into CellProfiler, module Images
At module NamesAndTypes, name the images accordingly, and set the type as "Color images"
Add a module called UnmixColors, use any “stain” that would extract the headlight beautifully.
Then add a module named ImageMath, use operation “Invert” on the extracted pictures from UnmixColors
Finally add module IdentifyPrimaryObjects, and use inverted image as input.

Hope that helps