Animated tap-to-focus reticle. Sits in the parent Stack; whenever position flips to a new non-null offset it
fades in (100 ms) then back out (300 ms). Renders the plugin's assets/focus_reticle.png.
Picks the active YOLO26 model size using a CupertinoSlidingSegmentedControl styled to match yolo-ios-app's
Main.storyboard modelSegmentedControl. Chips read YOLO26n / YOLO26s / YOLO26m / YOLO26l / YOLO26x; sizes not
yet on disk get a ↓ prefix to signal a download-on-tap. When downloadingSize matches a chip a thin
LinearProgressIndicator tracks downloadFraction under the label.
YOLO (You Only Look Once) is a class that provides machine learning inference
capabilities for object detection, segmentation, semantic segmentation, depth estimation, classification, pose estimation,
and oriented bounding box detection.
Real-time performance data for a YOLO inference: frame rate, processing
time, frame counter, and timestamp. Delivered via
YOLOView.onPerformanceMetrics.
The plugin uses method channels for communication between Flutter and native code.
Each platform (iOS, Android) provides its own implementation of the YOLO inference engine.
One-import camera UI matching the iOS showcase layout. Composes every widget under lib/widgets/, owns gestures
(pinch + tap-to-focus), drives the controller, and persists the last task across launches.
