Plant Species Classification
How did AiVirex build an AI that identifies India's medicinal plants at 95% accuracy?
AiVirex trained an ensemble of CNNs to 95% accuracy on 2,626 Indian medicinal plant species, with GPT-3 powering instant Q&A on each plant's uses.
The Numbers
Real results, measured
Case Study
The full story
AiVirex built an AI platform that detects and identifies indigenous Indian medicinal plants and returns their medicinal and culinary uses. It was a heavy R&D effort, large scale data gathering and preprocessing, then training across ResNet50, MobileNetV2, and VGG16 CNN architectures on 2,626 plant species spanning leaves, flowers, roots, stems, bark, and fruit. Stochastic weight averaging pushed accuracy to 95%, with custom model training and fine tuning throughout. GPT-3, enriched with web scraped data, powers instant Q&A about any identified plant.
Developed an AI driven platform to detect and recognize indigenous medicinal plants of India, providing details on their medicinal value and culinary usage. Trained on a dataset of 2,626 plant species (including leaves, flowers, roots, stems, bark, and fruits), the system utilized ResNet50, MobileNetV2, and VGG16 CNN models, achieving 95% accuracy through stochastic weight averaging. Integrated GPT-3 with enriched data from web scraping to enable instant Q&A about plants.
FAQ
Questions, answered
How accurate is AiVirex's plant identification model?
It reaches 95% accuracy across 2,626 species using an ensemble of ResNet50, MobileNetV2, and VGG16 with stochastic weight averaging.
Want results like this?
Tell us where you're stuck and we'll tell you what's actually possible, then scope it and give you a clear, tailored quote.