Questionário de IA em inspeção de qualidade

As respostas e explicações aparecem conforme você responde ao teste.

1. What is one practical difference between AI-assisted inspection and human inspection?
   • A model can apply a consistent learned or programmed check across many images, but it still needs validation and oversight
   • AI guarantees detection of every defect
   • Humans cannot inspect images
   • AI removes the need for quality standards
2. Por que a detecção de anomalias é útil quando os defeitos são raros?
   • Requer exemplos de todos os defeitos possíveis primeiro
   • Ele aprende como é o “normal” e sinaliza qualquer coisa que se desvie
   • Funciona apenas em texto, não em imagens
   • Ignora completamente eventos raros
3. What do the cited visual-inspection benchmarks evaluate?
   • Only OCR accuracy on scanned contracts
   • Speech diarization and transcription quality
   • Unsupervised anomaly detection and localization on MVTec datasets
   • The accuracy of a single human inspector
4. O que uma máscara de segmentação em nível de pixel informa ao inspetor?
   • O preço de venda do produto
   • Quão rápido o transportador está se movendo
   • O endereço da fábrica
   • Pixels that the model predicts as anomalous and that an inspector may need to verify
5. What deployment measurement should a factory team validate before placing an anomaly model on an inspection line?
   • Latency and throughput on the target hardware, image resolution, and pipeline configuration
   • A guaranteed zero-millisecond response on any device
   • A universal speed number independent of the line
   • A benchmark score that proves every production image is an anomaly
6. What challenging conditions does the MVTec AD 2 paper include in its benchmark?
   • Only clean, uniformly lit images with large defects
   • Examples include transparent or overlapping objects, lighting changes, and very small defects
   • Only audio recordings from industrial machines
   • Only synthetic images with no real photographs
7. What average AU-PRO did evaluated methods achieve on the MVTec AD 2 paper’s benchmark?
   • Exactly 100% for every method
   • Above 99% for all methods
   • Below 60% on average in the reported evaluation
   • The paper reports only speech word-error rates
8. Why can benchmark performance fail to transfer to a factory line?
   • Benchmarks always contain every defect type
   • The model cannot process pixels
   • Factory images are always identical
   • Real line conditions and defect distributions may differ from the benchmark dataset