<ul><li>DANCE is a proposed co-optimization approach for efficient segmentation model training and inference.</li><li>Current segmentation models suffer from expensive computation due to high-resolution images and multi-scale aggregation.</li><li>DANCE focuses on data-network co-optimization through input data manipulation and network architecture slimming.</li><li>Experiments show that DANCE achieves reduced training cost, less expensive inference, and improved mean Intersection-over-Union (mIoU).</li></ul>

DANCE: DAta-Network Co-optimization for Efficient Segmentation Model Training and Inference

Discover more