Computer Vision2020 – 2021ML Engineer
YOLO Autonomous Car Detection
Pretrained detector for 82 object classes
Object-detection model for autonomous driving assistance. Uses the YOLO algorithm, pretrained to recognise 82 object classes, intended to combine with sensor fusion for driving support.
The Problem & Engineering Constraint
The Core Challenge
A driving-assist stack needs real-time multi-class detection — not a single-label classifier — before it can be fused with other sensors.
Technical Architecture & Approach
Engineering Solution & Implementation
YOLO-based detector in TensorFlow / Keras with Matplotlib and SciPy around the training and evaluation loop. Positioned as an assist layer alongside sensor fusion.
Measured Production Impact
Verified Outcomes & Deliverables
82-class object recognition for driving scenes.
YOLO single-stage detector suitable for assistive autonomy work.
Technologies & Components
System Tooling & Technologies
PythonYOLOTensorFlowKerasNumPyMatplotlibSciPy