Journey to YUNO 02|Why Bowl Visibility Matters for Smarter Pet Feeding

A Simple Question Behind Smarter Feeding

Should a smart feeder dispense the next meal if there is still food left in the bowl?

Sometimes there is still food in the bowl when the next feeding time arrives. If fresh kibble keeps piling on top of leftovers, food stays exposed to air longer, which can gradually reduce freshness and crunch.

YUNO takes a smarter approach. Using its built-in camera and AI food recognition, YUNO checks whether food is still present before deciding to dispense. This helps avoid unnecessary feeding and makes automation better match the way pets actually eat.

Wide Enough to See. Clear Enough to Understand.

Reliable food recognition requires more than advanced AI. It also depends on choosing the right camera position, lens, and viewing angle.

The challenge was balancing two competing needs: the camera had to see enough of the feeding area to understand what was happening around the bowl, while still capturing enough detail for accurate food recognition.

A wider lens may sound like the easy answer, but it comes with trade-offs. Kibble appears smaller, edge distortion increases, and subtle details become harder to distinguish.

So instead of maximizing what the camera could see, we focused on maximizing what it could recognize.

During development, we tested nearly 30 camera configurations across different viewing angles, focal lengths, mounting heights, and structural designs. Each setup was evaluated not only for image quality, but for how reliably it could detect food in real feeding environments.

For AI-powered feeding, seeing the bowl is only the beginning. Understanding what is inside it is what really matters.

Prototype, Test, Rebuild

To find the right balance, we built and tested multiple 3D-printed prototypes.

Many early versions were far from beautiful. Exposed wires, visible gears, and unfinished printed parts made them look more like lab equipment than a consumer product. But every prototype taught us something.

The key question was simple: could YUNO keep its soft, rounded design while still giving the camera a clear, unobstructed view of the bowl?

Finding the answer took dozens of iterations. We adjusted the camera position, pan-tilt mechanism, food outlet, and bowl placement until the structure supported both the product design and the AI system behind it.

We also did not rely on renderings alone. After every change, we reviewed real footage from the camera itself to see exactly what the AI would see in a real feeding environment.

Real Camera Feed Testing

Prototype camera views helped us evaluate bowl coverage, blind spots, and the level of detail available for AI food detection.

The goal was not simply to fit the bowl into a wider frame. It was to keep the bowl clear enough for the AI to recognize what was inside.

A wider view means little without the image quality needed for reliable detection.

Over 80% Visible Bowl Coverage

After multiple rounds of testing and refinement, YUNO achieved over 80% visible bowl coverage in our internal validation while preserving the image quality required for reliable AI food detection.

And that is what bowl visibility means to us.

It is not just about seeing the bowl. It is not just about adding a camera. It is about giving YUNO the ability to recognize when food is already there, so it can make a smarter feeding decision before dispensing again.

YUNO checks first. Then it feeds smarter.