Links

Model

General Chain

• General Chain
  • FeaturePredictor
  • PickableSequentialChain

Feature Extraction

Feature extraction links extract feature(s) from given images.

• VGG
  • VGG16

Detection

Detection links share a common method predict() to detect objects in images. For more details, please read FasterRCNN.predict().

• Faster R-CNN
  • Detection Link
    • FasterRCNNVGG16
  • Utility
    • bbox2loc
    • FasterRCNN
    • generate_anchor_base
    • loc2bbox
    • ProposalCreator
    • RegionProposalNetwork
    • VGG16RoIHead
  • Train-only Utility
    • AnchorTargetCreator
    • FasterRCNNTrainChain
    • ProposalTargetCreator
• SSD (Single Shot Multibox Detector)
  • Detection Links
    • SSD300
    • SSD512
  • Utility
    • Multibox
    • MultiboxCoder
    • Normalize
    • SSD
    • VGG16
    • VGG16Extractor300
    • VGG16Extractor512
  • Train-only Utility
    • GradientScaling
    • multibox_loss
    • random_crop_with_bbox_constraints
    • random_distort
    • resize_with_random_interpolation

Semantic Segmentation

Semantic segmentation links share a common method predict() to conduct semantic segmentation of images. For more details, please read SegNetBasic.predict().

• SegNet
  • Semantic Segmentation Link
    • SegNetBasic

Classifiers

• Classifier
  • PixelwiseSoftmaxClassifier

Connection

• Connection
  • Conv2DActiv
  • Conv2DBNActiv