You Only Look Once: Unified, Real-Time Object Detection¶
Authors: Joseph Redmon, Santosh Divvala, Ross Girshick, Ali Farhadi Year: 2016 Citations: 30,000+
Summary¶
Single-stage object detector framing detection as regression. Predicts bounding boxes and class probabilities directly from full image in one forward pass. Enables real-time detection.
Key Concepts¶
- Single-Stage Detection: Direct regression without region proposals
- End-to-End Regression: Detection as regression problem
- Real-Time Performance: Fast enough for live video
- Spatial Awareness: Grid divides image for predictions
- Global Context: Entire image context aware
- Simplicity: Single CNN for direct output
Impact¶
- 30,000+ citations
- Introduced single-stage detection paradigm
- Enabled real-time object detection systems
- Faster than R-CNN variants
- YOLO v2, v3, v4, v5 widely deployed
- Influenced SSD and single-stage methods
- Critical for real-world applications
Related Papers¶
- R-CNN (Girshick et al., 2014)
- Faster R-CNN (Ren et al., 2015)
- SSD: Single Shot Detector (Liu et al., 2016)