Tanam.in

Tanam.in is a plant disease recognition Android app built as a team capstone project for the Bangkit Academy program (by Google, GoTo, Traveloka). The app helps gardeners and farmers identify plant diseases by photographing affected leaves, using a TensorFlow Lite model for on-device image classification. It also provides gardening education content and plant care recommendations.

Developed as a cross-functional team project with separate mobile, machine learning, and cloud computing tracks. The mobile team built the Android app using Kotlin with MVVM architecture. The ML team trained a convolutional neural network on plant disease datasets, then converted the model to TensorFlow Lite format for efficient on-device inference. UI/UX was prototyped in Figma before implementation to align the team on user flows.

• TensorFlow Lite model trained on plant disease image datasets, achieving 90-95% classification accuracy across multiple disease categories
• On-device inference enabling disease identification without network connectivity
• CameraX integration for capturing leaf photos directly within the app
• MVVM architecture with ViewModel and LiveData separating ML inference logic from the UI layer
• Gardening education content providing plant care tips and disease prevention guidance
• Team collaboration across three Bangkit tracks (mobile, ML, cloud) with coordinated API contracts

Completed and presented as the Bangkit Academy capstone project. The team demonstrated the app's disease recognition capabilities to program mentors and industry reviewers.