Product Overview

PoochGard AI

What You Can Do with PoochGard

Track Your Pet’s Health

  • View test history and trends over time

  • Secure, on-device storage with optional cloud sync

AI-Based LFT Analysis

  • Snap a photo of your test strip

  • Get results labeled as Positive, Negative, or Invalid

  • Detects even faint lines that humans may miss

Designed for Growth

  • Built for both pet owners and veterinary professionals

  • Subscription-ready with analytics tools for clinics

Real-Time Results

  • Instant feedback

  • Built-in concentration prediction using object detection + regression models

Cross-Platform Convenience

  • Available on iOS & Android

  • Lightweight AI models powered by YOLOv5 + TensorFlow Lite

Our Technology

Built on Flutter for cross-platform convenience, the PoochGard app uses cutting-edge technologies to power real-time diagnostics:

  • YOLOv5 Integration for object detection

  • TensorFlow Lite for on-device AI analysis & regression

  • Seamless UI for test history and user analytics

  • Subscription-ready infrastructure for clinics and pet parents

A feasibility study is currently underway to streamline this integration.

How It Works

1

Drop your sample on the kit

2

Take a photo using the PoochGard app

3

Let AI analyze the image

4

Get your result instantly

5

Save & review for future reference

Our AI Development Process

At PoochGard, we follow a 6-stage pipeline to ensure our AI-powered diagnostics are as reliable, accurate, and trustworthy as possible. Every step is designed to enhance learning quality and ensure top-tier performance in real-world conditions.

1. Sample Production

We begin by preparing specimen samples at various concentrations. These are applied to test kits to observe accurate color development for real-world simulation.

2. Data Collection

Using three or more smartphones, we capture images under 21 different color temperatures, producing over 63 images per sample, building a robust, diverse dataset.

3. Data Labeling

Images are manually labeled using an open-source tool to assign the correct results. This ensures high labeling accuracy, which is critical for AI learning.

4. Data Inspection

Our inspectors cross-verify concentration values and area ranges to validate labeling integrity and remove irrelevant or poor-quality data.

5. AI Training

With clean, labeled data, we train our models for:

  • Object detection – identifying key areas on the kit

  • Regression analysis – predicting concentration values

  • All models are then optimized for mobile efficiency.

6. AI Testing

We test with untrained (partitioned) data to measure generalization performance. If models perform well on unseen images, they’re considered ready.

Join the Future of Pet Health

Be a part of the growing community transforming pet care with PoochGard AI.