Get the latest tech news
Show HN: Real-Time YOLO Object Detection in Elixir: Fast, Simple, Extensible
A library for object detection and seamless YOLO model integration in Elixir - poeticoding/yolo_elixir
It works by predicting bounding boxes and class probabilities directly from an input image in a single evaluation, making it exceptionally fast compared to other object detection methods. The image is resized while preserving aspect ratio to fit within 640x640 pixels Any remaining space is padded with gray color (value 114) to reach exactly 640x640 This is handled by the FrameScaler behaviour and its implementations The image is converted to an Nx tensor with shape{1, 3, 640, 640} Pixel values are normalized from 0-255 to 0.0-1.0 range The channels are reordered from RGB to the model's expected format ( BGR in this case)
Or read this on Hacker News
