Screenshot 1
Screenshot 2
Screenshot 3
Screenshot 4
Screenshot 5
Screenshot 6
Screenshot 7
Screenshot 8
Screenshot 9
Screenshot 10
Screenshot 11
Mobile Application

LEAFAID AI - Plant Disease Detection App

AI-powered plant disease detection for farmers - scan, diagnose, and treat plant diseases instantly

Overview

The Problem

Farmers lose 20-40% of crops to diseases due to late detection and lack of expert knowledge.

The Solution

Mobile app with AI-powered image recognition for instant plant disease diagnosis, treatment recommendations, and agricultural marketplace.

Tech Stack

Frontend

Flutter (Dart) for cross-platform mobile

Backend

Firebase (Auth, Firestore, Storage, Cloud Functions)

AI/ML

Custom CNN (98-99% accuracy)
Google Gemini AI
TensorFlow Lite

Database

Firebase Firestore

DevOps

Firebase Hosting
Cloud Functions
Key Features
  • Real-time plant disease detection with 98-99% accuracy

  • Custom CNN ensemble model (EfficientNet + ViT + ConvNeXt)

  • On-device inference with TensorFlow Lite (<10MB model)

  • AI chatbot for plant care advice

  • Disease guide with 39+ plant diseases

  • Agricultural marketplace integration

  • Multi-language support for global farmers

Code Highlights
  • Ultimate ensemble model combining 3 architectures

  • Progressive training pipeline (multi-resolution)

  • Mobile optimization (quantization + pruning + knowledge distillation)

  • Mixed precision training for 50% faster training

  • Advanced augmentation (MixUp, CutMix, AutoAugment)

Impact & Results
  • 98-99% disease detection accuracy

  • <2 second inference time on mobile devices

  • 10,000+ scans processed

  • 90% model size reduction (100MB → <10MB)

  • Supports 39 plant disease classes

Architecture Note

Mobile-first AI application with on-device inference, custom CNN ensemble model, and real-time disease detection capabilities.

Mobile
AI/ML
Computer Vision
Flutter
TensorFlow Lite