On ImageNet validation, ResNet-101 RPN reduces one-crop oracle localization error using the ground-truth class from VGG-16's reported 33.1% to 13.3%, and dense multi-scale testing lowers it to 11.7%. With predicted classes, RPN gives 14.4% top-5 localization error, RPN plus R-CNN gives 10.6%, and an ensemble gives 8.9%.

On ImageNet validation, ResNet-101 RPN reduces one-crop oracle localization error using the ground-truth class from VGG-16's reported 33.1% to 13.3%, and dense multi-scale testing lowers it to 11.7%. With predicted classes, RPN gives 14.4% top-5 localization error, RPN plus R-CNN gives 10.6%, and an ensemble gives 8.9%. · CiteArk

Measurement results

0/6 assessed

0 supported0 disputed or conflicting6 not assessed

The complete result group stays with its claim. Assessed rows come first; their judgments do not replace the whole-claim conclusion. Expand a row for conditions, sources, and evidence history.

Each row is a measurement with its reported value, latest available observation, and cumulative evidence state.
Reported	Latest observation	Evidence judgment
33.1%	—	Not assessed
13.3%	—	Not assessed
11.7%	—	Not assessed
14.4%	—	Not assessed
10.6%	—	Not assessed
8.9%	—	Not assessed

Reported and observed measurements

Assessments (0)

Runs (0)

Measurement results