Computer Vision for Agriculture Hackathon
Transform the lab's dataset into impactful solutions! Join our 4-week innovation and collaboration event.

Dive into machine learning! Create a model with image datasets and take advantage of mentorship from industry leaders.

Computer Vision for Agriculture Hackathon
Makerere University
13 April - 17 May, 2024
Code. Create. Innovate.
The Computer Vision for Agriculture Hackathon 2024 is a unique event organized by Makerere AI Lab. This hackathon aims to provide undergraduate and postgraduate students from all universities in Uganda with the opportunity to work on computer vision tasks such as image classification and object detection. The participants will have the chance to work with some of the cutting-edge datasets collected under lacuna-funded projects.

The Hackathon will last four weeks, starting on April 13, 2024, and concluding on May 17, 2024. Throughout the event, participants will be engaged in various activities, including the official launch, where the necessary tools and tasks will be introduced; the main hackathon phase, which will include presentations and testing; and the closing ceremony, where the competition's winners will be announced.
Machine learning can solve societal issues in low- and middle-income regions, but unbiased, labeled datasets are needed for its effectiveness. The Lacuna Fund aims to bridge this gap by empowering researchers to create labeled datasets that tackle challenges in their communities. Accurate dataset labeling is crucial for the success of AI-enabled tools in agriculture, and the Lacuna Fund plays a crucial role in supporting initiatives to create, expand, and maintain training and evaluation datasets in this field.

The Computer Vision for Agriculture Hackathon 2024 aims to utilize several collected datasets in support of the Lacuna fund. These datasets include the Maize image datasets and the Cocoa datasets. The maize image datasets were collected by Makerere AI Lab in Uganda; the Nelson Mandela African Institution of Science and Technology in Tanzania, the Namibia University of Science and Technology in Namibia and KaraAgro AI in Ghana. The cocoa image datasets were collected by KaraAgro AI in Ghana. These datasets have been curated, labeled, and made publicly available to facilitate research, development, and innovation in the agricultural sector.
The hackathon winners will be determined by the models they submit and the mobile applications they develop. The hackathon organizers have created a model evaluation pipeline to help evaluate the submitted models' results.
Initiation of applications period
You can now apply with your team, or solo to be a part of the biggest AI Hackathon in Uganda. Explore further to find out about the challenges, mentors and prizes.
25 March, 2024
Official Launch
Toot- toot! Time's up and now the application process has emded, you will be notified if you have been selected to participate in the hackathon.the necessary tools and tasks will be introduced
13 April, 2024
Hackathon Phase
Have great workshops with the greatest minds of the market and start hacking!! This session will include presentations and testing of the deployed solutions.
13 April - 16 May, 2024
Closing Ceremony
This is the culmination of your hard work! It's time to acknowledge the teams who have translated ideas into the most well-executed, impactful solutions. Let's distribute the prizes and commend their outstanding efforts.
17 May, 2024
The participants will be divided into teams, and the teams' main purpose will be to develop two mobile applications powered by classification and detection models based on the Maize and Cocoa datasets.

Tasks for the Maize dataset


The classification models will be developed to classify the maize images into five different classes: healthy images, Maize Leaf Blight (MLB) disease, Maize Streak Virus (MSV) disease, Fall Army Worm (FAW) disease, and Maize Lethal Necrosis (MLN).


The detection models will be developed to detect some of the maize diseases from the annotated data by the experts. Three diseases were labeled in the data: MLB, MSV, and MLN. The developed models must be able to detect these three diseases.

Mobile Application

The classification and detection models will need to be deployed on a mobile application to facilitate farmers' use of these models in field setups.

Tasks for the Cocoa dataset


Cocoa stage detection models will have to be developed for the cocoa dataset. These models will have to detect the four stages of cocoa annotated by the experts: cocoa mature unripe, cocoa immature, cocoa ripped, and cocoa spoilt.

Mobile Application

A mobile application that can detect the cocoa stage using the best model will have to be developed for use in the field by farmers.
Meet the Organisers

Makerere AI Lab
Makerere AI Lab is a lab at Makerere University in Uganda, within the College of Computing and Information Sciences (CoCIS). The lab applies Artificial Intelligence and data science methodologies, including Computer Vision and Natural Language Processing, to tackle challenges in developing countries. The lab aims to advance AI research for practical solutions, focusing on disease diagnosis, auction design, traffic pattern analysis, and infectious disease prediction. Their vision is to strive for excellence and to create accessible AI solutions that significantly impact communities and contribute to sustainable development.
Follow on twitter, facebook and instagram to get the latest news about the hackathon.