Computer Vision
  • Introduction
    • Neural network basics
    • MLE and Cross Entropy
    • Convolution basics
    • Neural Network Categories
  • 2D Backbones
    • ResNet
    • Transformer
      • Recurrent Neural Network
      • Vision Transformer
      • SwinTransformer
  • Methods for Object Detection
  • Object Detection
    • The R-CNN family
    • ROI pool & ROI align
    • FCOS
    • Object Detection in Detectron2
  • Segmentation
    • Fully Convolutional Network
    • Unet: image segmentation
  • Video Understanding
    • I3D: video understanding
    • Slowfast: video recognition
    • ActionFormer: temporal action localization
  • Generative models
    • Autoregressive model
    • Variational Auto-Encoder
    • Generative Adversarial Network
    • Diffusion Models
    • 3D Face Reconstruction
Powered by GitBook
On this page
  • 1. Use Feature Pyramid to extract features
  • 2. Use Focal loss to handle imbalance classes
  • 3. Use three separate losses to train the network
  1. Object Detection

FCOS

An anchor-free single stage object detector

PreviousROI pool & ROI alignNextObject Detection in Detectron2

Last updated 2 years ago

1. Use Feature Pyramid to extract features

2. Use Focal loss to handle imbalance classes

3. Use three separate losses to train the network

Overall model structure
Use feature pyramid to get image features at different scales. 1) Use top down connection to provide lower layers with high level features. 2) Efficient because all levels use the same backbone and