Object detection intro — YOLO overview — Deep Learning & Neural Networks

Image classification tells you what is in an image. Object detection tells you what is in the image and where — outputting both a class label and a bounding box (x, y, width, height) for every detected object. Detection is far more useful for real applications: a cricket broadcast system does not just need to know 'there is a ball somewhere' — it needs the ball's exact pixel coordinates to overlay trajectory graphics, track speed, and trigger DRS reviews. YOLO (You Only Look Once), introduced by Redmon et al. in 2015 and now in its 8th major version (YOLOv8 by Ultralytics), is the dominant real-time object detection architecture because it frames detection as a single regression problem: instead of a two-stage pipeline (first propose regions, then classify them), YOLO divides the image into a grid and simultaneously predicts bounding boxes and class probabilities for all grid cells in one forward pass. This makes YOLO 10–100x faster than two-stage detectors like Faster R-CNN while achieving competitive accuracy. Understanding the YOLO paradigm — grid-based prediction, anchor boxes, IoU-based loss, non-maximum suppression — gives you the vocabulary to use, fine-tune, and evaluate any modern detection system.

Analogy🏏Cricket

🏏 Think of it like cricket: The Duckworth-Lewis-Stern (DLS) method, the ICC bowling economy metric, and the net run rate each measure team performance but optimise for completely different things — DLS cares about wickets-in-hand, economy cares about runs-per-over, NRR cares about tournament-wide margins. Using the wrong metric to evaluate a bowler is like using MSE for classification: technically computable but optimising the wrong thing entirely. Just as you would never rank a spinner by batting average, you should never use MSE when your model predicts probabilities. Just as focal DLS adjustments down-weight easy chases and amplify close finishes, focal loss down-weights easy examples and amplifies hard minority-class examples. The choice of scoring metric defines what excellence means — and so does the choice of loss function.