INIS
neural networks
100%
images
100%
defects
100%
humans
80%
validation
80%
datasets
60%
performance
60%
pipes
60%
classification
60%
labor
40%
equipment
20%
sensitivity
20%
levels
20%
precision
20%
specificity
20%
testing
20%
metrics
20%
interventions
20%
porous materials
20%
evaluation
20%
connections
20%
machine learning
20%
leakage
20%
inspection
20%
recommendations
20%
errors
20%
reviews
20%
data
20%
Computer Science
Convolutional Neural Network
100%
Classifier
40%
Human Operator
40%
Classification Performance
20%
Large Data Set
20%
Overestimation
20%
Testing
20%
Research Worker
20%
Machine Learning Technique
20%
Evaluation Metric
20%
Precision
20%
Classification Task
20%
Keyphrases
Common Defects
20%
Human Labor
20%
Porous pipe
20%
Automated Classifier
20%
Immunology and Microbiology