Can you describe the project that you have been working on?
‘Certainly! The project was in collaboration with the Netherlands Plant Eco-phenotyping Centre (NPEC). They tasked us with measuring the primary root of plants grown in a petri dish. We did so by using pictures taken by an advanced camera, based on those pictures we measured the length of the roots. The goal was to measure and precisely locate and moisten the root tips.’
‘My fellow students and I labelled and selected over a hundred root images collectively, fed the data into a neural network, and fine-tuned it to identify and measure roots accurately. A Kaggle competition added a competitive edge, pushing us to optimise our models further.’
What was your role in this project?
‘As this was an individual project, I took on several roles. The work involved preparing the images, eliminating unnecessary elements, and optimising them for efficiency. I played the part of an AI engineer, experimenting with various parameters to train the neural network. After the training, my focus shifted to measuring the root length, relying more on coding than AI.’
‘During the robot training phase, we collaborated in groups, tweaking parameters for optimal results. The final deliverable was a highly accurate model integrated into the robot, capable of nano-meter precision. I am proud to say that I finished in the top five of the Kaggle competition.’
What are your future plans?
Wesley: ‘One of the standout aspects is the freedom to explore and apply our own creative ideas. The curriculum guides us in the right direction, but it’s up to us to bring our unique perspective into play.’
Feel free to connect with Wesley on LinkedIn for more insights into his projects and journey in the field of Data Science & AI.
