PaperCVPR · 2015

Fully Convolutional Networks for Semantic Segmentation

Jonathan Long Evan Shelhamer Trevor Darrell

arXiv 1411.4038 DOI View in graph

1page built on it

9more pages cite it

4narratives

18registry papers cite it

In the Atlas

ModelPrimary source of

FCN: Fully Convolutional Networks

Encoder-decoder CNN for dense pixel-wise classification — converts ImageNet classifiers into fully convolutional networks via 1×1-conv reinterpretation, then upsamples via learnable bilinear-initialised deconvolution with skip connections from earlier pooling stages.

Cited as a reference by 9 pages, grouped by domain

Segmentation 6

BiSeNet DeepLab Fast-SCNN Mask R-CNN SegFormer U-Net

Features 2

MobileNetV3 ResNet

Detection 1

Faster R-CNN

Appears in 4 narratives

NarrativeDetection, Removing the Machinery NarrativeSegmentation, from Energy to Prompt NarrativeVision Under a Latency Budget NarrativeWhere Did the Inductive Bias Go?

Lineage

Papers in the Atlas registry only · one dot per paper

Built upon by · 18

Builds on · 0

Built upon by · 18

2014Very Deep Convolutional Networks for Large-Scale Image RecognitionSimonyan, Zisserman 2015Faster R-CNN: Towards Real-Time Object Detection with Region Proposal NetworksRen, He et al.2015U-Net: Convolutional Networks for Biomedical Image SegmentationRonneberger, Fischer, Brox 2016Deep Residual Learning for Image RecognitionHe, Zhang et al.2017Mask R-CNNHe, Gkioxari et al.2017Feature Pyramid Networks for Object DetectionLin, Dollár et al.2017Focal Loss for Dense Object DetectionLin, Goyal et al.