YOLO11 launched as latest SOTA vision model. Supports object detection, segmentation, and classification. Available in PyTorch, ONNX, CoreML, and TFLite formats
Ultralytics offers various models for object detection, segmentation, and classification. YOLOv3-11 versions provide real-time object detection capabilities. SAM, SAM2, MobileSAM, and FastSAM models support segmentation tasks. YOLO-NAS and RT-DETR models offer real-time detection capabilities
Object detection models identify and classify relevant objects in images. Models predict bounding boxes and assign classes to them. Models can tradeoff precision for recall by adjusting confidence thresholds
ESP32-CAM is a 9-dollar board capable of simple object detection. Object detection combines classification and localization technologies. TinyML enables machine learning on low-power microcontrollers
YOLOv8 is Ultralytics' latest object detection model using PyTorch. Model performs better than YOLOv5 64% of the time on RF100 dataset. Supports object detection, instance segmentation, and image classification
OpenCV DNN module enables deep learning inference on images and videos. Module is highly optimized for Intel processors. Supports various deep learning frameworks including Caffe, TensorFlow, Torch, and PyTorch