You Only Look Once: Unified, Real-Time Object…

YOLO asks how a computer can locate and name every object in an image quickly enough for live video. Read it to see how one full-image network fills a fixed responsibility grid with boxes and class scores, why that unified prediction runs in real time, and where coarse spatial capacity trades localization accuracy for speed.

Reading focus: How the cell containing an object's center takes responsibility and turns a variable scene into a fixed 7 × 7 × 30 prediction tensor on PASCAL VOC. Why class-specific confidence combines object presence, predicted-versus-truth overlap, and conditional class probability before thresholding and duplicate removal. What the paper's VOC speed, mAP, localization, background-error, small-object, objective, and historical-hardware limits do and do not establish.

CVPR 2016. Joseph Redmon, Santosh Divvala, Ross Girshick, and Ali Farhadi. 30 min read, easy difficulty.