Introduction

Plant AI is an innovative web application designed to identify and diagnose plant diseases at an early stage, predominantly in crop leaves. Developed with the primary objective of assisting farmers, this AI-powered tool is not only highly accessible and user-friendly but also operates seamlessly across various devices, including low-end phones, both online and offline.

Project Inception

The Genesis

In 2021, amidst my journey in the realm of development and unbeknownst to me, a career in product design, I participated in a hackathon organized by Microsoft. My foray into this event was propelled by an invitation from a friend I met in the Microsoft Student Community. Recognizing my flair for design, he envisioned my contribution as pivotal to the project. Thus, our trio embarked on a mission to build what would later evolve into Plant AI during my third year of college.

The Challenge

Tasked with the goal of creating a prototype that would enable farmers to detect and manage plant diseases promptly, our team aimed to blend innovation with practicality, ensuring the solution was accessible and beneficial to the end user.

The Team Dynamics

Our team comprised three budding talents in their early bachelor's years:

• Rishit Dagli - Spearheaded Machine Learning.

• Rucha Yagnik - Mastered API Integration & Deployment.

• Rishabh Singh - Led Designing & Frontend Development.

My Role and Workflow

Functioning in a dynamic, non-linear fashion, our team adopted an asynchronous mode of collaboration, leveraging Google Meet to connect and progress. My responsibilities were multi-faceted, encompassing wireframe design, React web app development, and CSS crafting. Furthermore, I ventured into promotional realms, designing banners and presentations for various events to showcase Plant AI.

Development Journey

The Machine Learning Model

Our AI model, trained on over 87,000 images of both healthy and diseased plant leaves, utilized an open-source dataset. Interestingly, the application transcended its initial agricultural scope, finding relevance and utility in everyday life scenarios.

Platform Decision-Making

Confronted with choices between a mobile app, a web app, and a hardware-based solution, we opted for a web app. This decision was driven by the need for a swift, installation-free Proof of Concept and the imperative to optimize the model for portability, ensuring functionality even on low-end smartphones without GPUs.

Crafting the Web App

From Wireframes to Reality

In the realms of wireframe design, I embraced the learning curve of Figma, bringing the envisioned design to life with meticulous CSS. Despite initial hesitance with React, the final product resonated closely with our wireframes, a testament to our collaborative synergy, especially with Rucha's proficiency in integrating React and Machine Learning.

Deployment and Hosting

Our journey from development to deployment saw the Machine Learning model hosted on TensorFlow and the web app on Azure, with Rucha steering this critical phase.

Marketing and Community Engagement

Venturing into the uncharted territory of marketing, our team employed platforms like Reddit, Twitter, LinkedIn, and HackerNews to disseminate our project. The response, fueled by the support of the Microsoft Student Community and personal networks, was heartening. Feedback, particularly on fine-tuning the AI and advocating for open-source code, was invaluable.

Milestones and Reflections

Achievements

Plant AI also achieved quite a lot:

• Triumphed at Green-a-thon.

• Secured victory at Microsoft’s Shark Tank Event.

• Featured at the Global AI Student Conference.

• Showcased in an interview and demo on Microsoft's “Hello World” Live.

Learnings and Revelations

The project was not just a technological venture but a holistic learning experience. It underscored the essence of teamwork, the art of presentation, and the pivotal role of marketing even in a hackathon setting. Reflecting on our journey, we acknowledge the potential for a more streamlined process and a heightened focus on marketing efforts.

Conclusion

The journey of Plant AI, from a mere concept to a tangible tool empowering farmers, was a blend of innovation, learning, and invaluable teamwork. It underscored the potential of technology to transform traditional sectors and the importance of a multidisciplinary approach in bringing visions to life. As we look back, the experience remains a testament to the power of collaboration, perseverance, and the relentless pursuit of innovation.