A graph-based approach to image segmentation uses an object’s bounding box to transform image pixels into a weighted graph. A shortest-path search then traces a closed contour, with reported tests demonstrating its use on vehicle images and other object classes.
Articles Tagged: computer vision
Traffic Sign Detection: A Two-Stage Vision System
A two-stage computer vision system detects traffic sign candidates using color and shape, then classifies them with a convolutional neural network. Its design highlights practical trade-offs in preprocessing, recognition accuracy and operating conditions.
5G Robotics: Scaling AI for Industrial Automation
Combining 5G connectivity with on-device AI can help robotics developers address demanding requirements for autonomy, sensing and communication. Learn how an integrated robotics platform brings computing, vision, connectivity and development tools together.