Normalized Cut Loss for Weakly-Supervised CNN Segmentation

By Meng Tang, Abdelaziz Djelouah, Federico Perazzi, Yuri Boykov, Christopher Schroers. 2018

346 citations recorded by the source corpus.

Most recent semantic segmentation methods train deep convolutional neural networks with fully annotated masks requiring pixel-accuracy for good quality training. Common weakly-supervised approaches generate full masks from partial input (e.g. scribbles or seeds) using standard interactive segmentation methods as preprocessing. But, errors in such masks result in poorer training since standard loss functions (e.g. cross-entropy) do not distinguish seeds from potentially mislabeled other pixels. Inspired by the general ideas in semi-supervised learning, we address these problems via a new principled loss function evaluating network output with criteria standard in "shallow" segmentation, e.g. normalized cut. Unlike prior work, the cross entropy part of our loss evaluates only seeds where labels are known while normalized cut softly evaluates consistency of all pixels. We focus on normalized cut loss where dense Gaussian kernel is efficiently implemented in linear time by fast Bilateral filtering. Our normalized cut loss approach to segmentation brings the quality of weakly-supervised training significantly closer to fully supervised methods.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Computer Vision and Pattern Recognition

Citations in the corpus

Incremental Learning in Semantic Segmentation from Image Labels (2021, methodology): “To address this issue, Weakly Supervised Semantic Segmentation (WSSS) methods aim to obtain effective segmentation models using cheaper supervisions such as bounding boxes [18, 32, 49], scribbles [39, 60], points [6, 53], and image-level labels [31, 34, 35, 59].”
SG-One: Similarity Guidance Network for One-Shot Semantic Segmentation (2018, neutral): “[8] adopted spectral”
Pseudo-mask Matters in Weakly-supervised Semantic Segmentation (2021, methodology): “To reduce the annotation burden, weaklysupervised semantic segmentation approaches have been proposed using scribble annotations [23, 32], bounding boxes [36, 9, 16] , points [3] or image-level labels [17, 27, 28, 2, 33].”
C-MIDN: Coupled Multiple Instance Detection Network With Segmentation Guidance for Weakly Supervised Object Detection (2019, background): “Weakly Supervised Semantic Segmentation A signiﬁ-cant advance of performance of WSSS has been witnessed in last several years [44, 16, 38, 1, 29, 45, 41].”
On Regularized Losses for Weakly-supervised CNN Segmentation (2018, methodology): “This continues the study of losses motivated by standard shallow segmentation energies started in [3] with normalized cut loss.”
BigDatasetGAN: Synthesizing ImageNet with Pixel-wise Annotations (2022, background): “Reducing labeling costs can be achieved in a variety of ways, such as interactive humanin-the-loop annotation [2, 50, 53, 64, 73], nearest-neighbor mask transfer [26, 51, 63, 70], or using weak supervision in the form of cheaper labels such as object boxes [7, 43, 45], scribbles [48, 69] or very coa”
Interactive Object Segmentation With Inside-Outside Guidance (2020, methodology): “Particularly, image-level labels [50, 61, 63, 62, 27, 32], points [3, 51], bounding boxes [18, 33] and scribbles [39, 58, 59] have been employed as guidance to supervise the training of semantic segmentation networks.”
Interactive Object Segmentation With Inside-Outside Guidance (2022, methodology): “Particularly, image-level labels [67], [68], [69], [70], [71], [72], points [73], [74], bounding boxes [75], [76], scribbles [77], [78], [79] have been employed as guidance to supervise the training of semantic segmentation networks.”