← Back to Projects Segmentation using YOLO
Segmentation pipeline with YOLO models and optimized inference exports.
Categories
CVML
Tech Used
YOLOv8PyTorchsegmentation modelsfine-tuningOpenCVUltralyticsONNXTensorRTCUDADocker
Problem
Pixel-level segmentation needed for measurement/inspection tasks.
Approach
- Trained YOLO segmentation model with augmentation
- Improved dataset quality and evaluated IoU/precision/recall
- Exported to ONNX/TensorRT for faster inference
Results
- Production-ready segmentation
- Optimized inference path
Demo Videos