With Latitude AI, Break Through Tech Fellows explore how cutting-edge computer vision powers the next generation of self-driving cars—and how artificial intelligence can improve how vehicles perceive their environment.
Published
06/25/2025
How can we help cars see the world in 3D? That’s the question behind a recent AI Studio challenge project with Latitude AI, a company at the forefront of autonomous driving innovation. The project invited Break Through Tech AI Fellows to explore how cutting-edge computer vision powers the next generation of self-driving cars—and how artificial intelligence can improve how vehicles perceive their environment.
The Challenge: Train a Car to Understand Its Surroundings
Self-driving cars rely on a suite of sensors, cameras, and algorithms to navigate the world. But raw data isn’t enough—autonomous systems must make sense of everything they “see” in real time. That’s where 3D object detection comes in.
In this challenge, fellows were tasked with building a 3D object detection model that could identify and classify objects, like cars, pedestrians, and traffic signs, from images. Unlike traditional 2D detection, this project focused on helping the system interpret depth and perspective, mimicking how real autonomous vehicles detect spatial relationships between objects on the road.
The Approach: From Pixels to Perception
AI Fellows dove into the world of deep learning and computer vision, training convolutional neural networks (CNNs) on publicly available datasets to recognize common road objects. They experimented with transfer learning to boost performance and explored innovative techniques to transform images into a bird’s eye view (BEV)—a critical step in autonomous driving systems that simplifies spatial understanding.
The team also investigated ways to represent objects in 3D using their visible edges and corners, pushing beyond simple bounding boxes. This helped the models understand not just where an object is, but how it exists in space—a key challenge in building safer autonomous navigation systems.
Real-World Impact: Shaping the Future of Mobility
The implications of this work go far beyond the classroom. Accurate 3D object detection is vital for the safety and efficiency of autonomous vehicles, enabling them to make smarter, faster decisions on the road. By working on this challenge, fellows contributed to a field with the potential to reduce accidents, ease congestion, and transform the way we move.
The Takeaway
For the fellows in this AI Studio challenge, the journey was about more than just building a model. It was about understanding how artificial intelligence can see the world, and how that vision can make driving safer for everyone.
Want to build AI for a better world? Learn more about Break Through Tech’s AI Studio and the challenge projects helping undergraduate students across the country turn classroom learning into real-world impact.
Break Through Tech fellows bring extensive training and fresh insights to complex business challenges. Learn more about how AI Studio can support your next project and offer you a first look at exceptional early tech talent.