Graphic to accompany blog post on new AI Nodes

[Steve-Mcl]

Related Issue

No response

Description

Considerations:

• Where will the image be displayed/shared, product? blog post? social media? presentation? website?
  • To be used as the main title image in a new blog post

Inspirations:

• Something that depicts AL/ML inference of data in, to something out
  • e.g. An image in --> A text classification out
  • e.g. Text in --> Speech out
• ideas:
  • https://www.aimtechnologies.co/wp-content/uploads/2023/06/AI-Image-Recognition.jpg
  • https://fritz.ai/wp-content/uploads/2023/09/how-image-recognition-works-.jpeg
  • https://www.folio3.ai/blog/wp-content/uploads/2022/03/Text-To-Speech.jpg
  • This AI generated mess has some ideas (The depiction of pipelines to signal data in ~ results out is a good and transferable metaphor for AI/ML operations)
  

[Steve-Mcl]

@Yndira-E - any ETA on this please?

[Yndira-E]

Hey @Steve-Mcl, it'll be done tomorrow.

[Steve-Mcl]

Tomorrow is perfect. Thanks Yndira.

[Yndira-E]

Hi @Steve-Mcl Would this make sense?

[Steve-Mcl]

A good start but the title... I dislike that it wraps - thought TBH, I cannot thing of anything that would fit on one line!

Perhaps "FlowFuse AI Nodes: Build AI flows in low code" or "Bring Your ONNX Models to Node-RED with FlowFuse AI Nodes" or "Run Custom AI Models in Node-RED with FlowFuse AI Nodes"
Oh crumbs, those ideas are probably longer again 😆

Gemini suggested the input should be more generic as if to say "unknown thing going in" with "known thing coming out on the right" - it also adds a star on the lower right (cute :D )

Is my pathetic attempt at feeding back useful info regarding graphic design of any use at all 🤔

[Yndira-E]

Thanks @Steve-Mcl that's useful feedback!

The title usually matched the article title, so let me know if you'll update it too, which option you go for, and I'll update it.

Regarding the graphics. I understood that the training and recognition were based on images. Is that the case, or did I get it wrong? Thinking that the objects going into the funnel could be polaroids or JPEG icons, if that's the case.

I see the arrows going outwards from the circle, is that correct? Or just AI being AI?

And what do the icons around the circle mean in this flow? What are they representing, floating around like that?

And the "result" would be it recognized a single thing or multiple at once?

[Steve-Mcl]

Hi @Yndira-E - I know we talked this over on slack, but for those following along:

The new AI Nodes can use images but AI is not just images. The ONNX models take input data (in this case image is the data).but that could equally have been 1+0s and getting 0+1 in the result (not not great content for human consumption)
For example, another one of the built in demos that come with the nodes is converting Text To Speech (no images were hurt in the making of that example flow) - however, for the article, I decided to work on training an image classifier, because it was easy to do, highly visual and quite impactful (and something manufacturing can use e.g. point a camera at a production line and classify their own widgets for decanting or redirecting on a set of conveyors)

[Yndira-E]

@Steve-Mcl how about now?

[Yndira-E]

@Steve-Mcl here's the final one