Idanwo ti o sopọ mọ itọsọna · Alabọde Ipele

Ekun-Da CNNs adanwo

Test region proposals, shared convolutional features, RoI operations, masks, and the accuracy-speed tradeoffs of two-stage detectors.

Jẹmọ itọsọna awọn ọnaEkun Da Cnns
Ibeere 1 ti 8

Kini imọran pataki ti CNN ti agbegbe kan?

Awotẹlẹ gbogbo ibeere8 ibeere

Awọn idahun ati awọn alaye han bi o ṣe n mu ibeere naa.

  1. Kini imọran pataki ti CNN ti agbegbe kan?
    • Dabaa awọn agbegbe oludije, lẹhinna ṣe lẹtọ ati awọn nkan apoti ni ọkọọkan
    • Ṣẹda aami ẹyọkan fun gbogbo aworan naa
    • Tẹ aworan naa ṣaaju ṣiṣe
    • Yi aworan pada si iwọn grẹy ni akọkọ
  2. Why was the original R-CNN computationally expensive?
    • It trained without image pixels
    • It ran a CNN separately on each proposed region
    • It required one human label per test image
    • It computed only one shared feature map
  3. Kini Yiyara R-CNN ṣafihan lati rọpo Wiwa Yiyan?
    • A o tobi CNN ẹhin
    • Igbesẹ iyipada grẹy kan
    • Nẹtiwọọki Igbero Agbegbe kan (RPN)
    • Nẹtiwọọki ti nwaye loorekoore
  4. Kí ni RoI pooling ṣe?
    • O npa awọn agbegbe pẹlu awọn ikun kekere
    • O mu ipinnu aworan pọ si
    • O dapọ gbogbo awọn agbegbe sinu ọkan
    • O gbin ati ṣe atunṣe awọn ẹya lati maapu ẹya ti o pin fun agbegbe kọọkan
  5. Agbara afikun wo ni Mask R-CNN ṣafikun?
    • Awọn iboju iparada ipin ipele Pixel fun ohun kọọkan ti a rii
    • Igbasilẹ ohun
    • Itumọ ọrọ
    • Ijinle 3D lati ẹbun kan
  6. What training objective does Mask R-CNN use for its classification, box, and mask branches?
    • Only a pixelwise mask loss with no classification
    • A multi-task loss that combines classification, bounding-box regression, and mask losses
    • A language-model next-token loss
    • A region-proposal loss with no object classification
  7. What alignment problem does RoIAlign address compared with RoI pooling?
    • It replaces the Region Proposal Network with Selective Search
    • It removes all need for a shared feature map
    • It avoids the coordinate quantization that can misalign extracted region features
    • It predicts a class label before processing the image
  8. How should two-stage and single-stage detector speed/accuracy be compared?
    • Two-stage models are always more accurate and faster
    • Single-stage models cannot detect multiple objects
    • The architectures have identical pipelines
    • Using a specified benchmark and implementation, because the tradeoff is not universal