Documentation Application Tour Get Involved

Object Detection

Simulated iPad object inspector — Figure 1. iPad — Detection label, confidence, evidence, and human review state. Simulated research data.

Object detection proposes a category and rectangular image region. It does not identify a specific real-world individual and it can miss, duplicate, or misclassify objects. OrganicVision preserves alternate labels and original model confidence because confidence scales differ between models.

Model-specific labels pass through a versioned mapping into the Phase 5 taxonomy. Unmapped labels are retained as unmapped:<label> and require review. Live messages are bounded to 50 observations and batched; large vectors and archives do not use the preview channel.

 _   _           _     ____
| | | | ___ _ __| |__ |  _ \  _____   __
| |_| |/ _ \ '__| '_ \| | | |/ _ \ \ / /
|  _  |  __/ |  | |_) | |_| |  __/\ V /
|_| |_|\___|_|  |_.__/|____/ \___| \_/