09-14, 11:20–11:50 (Europe/Amsterdam), Bar
In the world of computer vision, the focus is often on cutting-edge neural network architectures. However, the true impact usually lies in designing a robust system around the model to solve real-world business challenges. In this talk, we guide you through the process of building practical computer vision pipelines that leverage techniques such as segmentation, classification, and object tracking, demonstrated by our predictive maintenance application at Port of Rotterdam. Whether you're an experienced expert seeking production-worthy pipelines or a novice with a background in data science or engineering eager to dive into image and video processing, we will explore the use of open-source tools to develop and deploy computer vision applications.
This talk provides a comprehensive demonstration of creating a powerful computer vision pipeline using widely-used libraries such as PyTorch, torchvision, and OpenCV. We break the pipeline down into manageable components, discussing the importance of proper separation of concerns. Onboarding new use cases becomes a breeze when following best practices in the project structure, combined with user-friendly command-line interfaces. Efficient development and validation processes are ensured by designing a sane data model and writing useful tests. Additionally, we explore the critical topic of maintainability, applying MLOps principles for long-term success.
To bring these concepts to life, we present a real-world application: the Machine Learning Inspector. This predictive maintenance tool, deployed at the Port of Rotterdam, automatically detects and inspects objects in video streams from trucks and ships, delivering actionable insights. We discuss how we work together with asset inspectors to capture their knowledge of the real world in our artificially intelligent computer vision tool.
Join us in this talk to gain practical knowledge and valuable insights for designing, deploying, and maintaining computer vision pipelines that drive tangible impact. We aim to empower the audience to build their own computer vision pipelines; with the right design philosophy, every data professional should be able to build computer vision pipelines that might be complex, but not complicated.
No previous knowledge expected
Wesley is the Lead Machine Learning Engineer at Xccelerated (part of Xebia). There, he trains and guides junior-to-medior ML Engineers in Xccelerated's one-year program. Besides that, he works as an MLE on various projects, recently at KLM, ProRail, and Port of Rotterdam. In his free time, he likes to stay up-to-date in the ML ecosystem and play around with computer vision.
Jeroen is an expert in machine learning and AI, specializing in transforming ideas and proof-of-concepts into value-driven products. Leveraging deep expertise in data science and engineering, he offers practical solutions to enhance machine learning infrastructure and elevate data teams' AI skills.