Unity AR Animal Detection App

I need a Unity-based Android application that can spot and analyse both domestic and wild animals in real time. Shape, colour and overall body type will feed an AI model so the system can tell the user what they are looking at, then—through augmented-reality overlays—display the animal’s estimated age and distance, its breed or species, and even its common local name (for example, “camel wadhah” when a white camel is recognised).

Augmented-reality precision is the single most important part of this project, so the AR layer must stay stable across a wide range of Android handsets and lighting conditions. A clean, visually pleasing interface should let users tweak detection thresholds or confidence values without digging through menus; the experience needs to feel smooth for both casual and professional use.

Core scope
• Unity build for Android (latest LTS version preferred)
• Computer-vision / AI pipeline that runs on-device, with fallback to cloud inference if you think it is necessary
• AR Foundation or similar framework to place age-and-distance readouts directly on the camera feed
• Dynamic UI that follows modern Material guidelines and is easily reskinned
• Thorough testing on a representative device matrix so the app behaves consistently across manufacturers

Acceptance criteria
1. The camera feed correctly identifies at least five domestic and five wild species in a controlled demo video.
2. Age and distance overlays remain locked to the target while the user moves the handset within a three-metre radius.
3. Breed/species and local-name strings appear in less than two seconds after initial detection.
4. Tweakable parameters (confidence slider, colour tolerance, etc.) can be saved and re-loaded without restarting the app.
5. A signed APK plus full Unity project are delivered, ready for onward store submission.

If you have prior work combining computer vision, AI and Unity AR on Android, please mention it in your bid and point me to a live sample or short demo clip.