Which AYNIcam AI model should you choose?

A model is not the same as a view

The AI model processes the camera image. A view decides how the app presents the resulting information. With the detector family, Detection, Maps, Virtual and Parking offer different presentations without loading four independent cameras. The segmentation family exposes different outputs and opens a dedicated page.

nawnaw: the Free detector

nawnaw is the Free detector based on YOLOX-Tiny. It supports person, bicycle, motorcycle, car, bus, traffic light and stop sign. Start here when you want detection boxes or one of the four driving views. A simpler model is not a guarantee of a particular scan rate, but it is a sensible starting point for comparing workload on your phone.

nawnaw pro: another detector option

nawnaw pro uses the larger YOLOX-Small detector and requires Pro. It shares the same seven categories and detector-family views. Try it when you want to compare detection results in your conditions. Do not read the model picker’s illustrative confidence numbers as a benchmark, or assume that a Pro detection is automatically correct.

nawnaw seg pro: masks and road surface

Choose this Pro model when you want instance masks and the road-surface output. It opens Segmentation rather than the detector family’s four pages. It does not add a building detector, lane-count recognition or reliable front/rear vehicle pose. If you want Maps or Parking again, return to nawnaw or nawnaw pro.

Make a fair comparison

While safely stationary, keep the mount and scene unchanged and switch one model at a time. Compare useful observations, missed objects and readability as well as responsiveness after warm-up. The three-card picker contains built-in illustrative previews; these help explain output types, not prove accuracy. Choose the output you need within your own device’s thermal budget.