SuperParsing: Scalable Nonparametric Image Parsing with Superpixels

Tighe, Joseph; Lazebnik, Svetlana

doi:10.1007/978-3-642-15555-0_26

Cited by 429 publications

(627 citation statements)

References 19 publications

Supporting

Mentioning

622

Contrasting

Unclassified

Order By: Relevance

“…Tighe and Lazebnik's SuperParsing algorithm [34] for scene parsing takes a similar nonparametric approach but operates on the level of superpixels. The query image's superpixels are labelled using a Markov random field model, based on similar superpixels in the query's nearest neighbor images in the database.…”

Section: Related Workmentioning

confidence: 99%

“…The retrieval set aims to find a subset of database images that are contextually similar to the query image, and is a typical component of nonparametric scene analysis methods [15], [21], [34]. In addition to filtering out semantically irrelevant database images that are likely to be unhelpful, a small retrieval set makes nearest neighbor based label transfer practical on large datasets.…”

Section: Forming the Retrieval Setmentioning

confidence: 99%

“…For θ we adopt the same pairwise potential term as in SuperParsing, which is based on the co-occurrences of category labels [34]: Fig. 4.…”

Section: Performing Mrf Inferencementioning

confidence: 99%

“…In recent years, the growth of online image collections and the adoption of crowdsourcing methods for annotating datasets have led to an interest in developing scalable methods that are suitable for open-universe datasets [34]. An open-universe dataset is one that is continually changing as users contribute new images and annotations, such as LabelMe [29].…”

Section: Introductionmentioning

confidence: 99%

“…Conceptually, superpixel based techniques reason about pieces of objects and apply auxiliary techniques on top to combine these pieces in a principled way. For instance, semantic label co-occurrence probabilities are commonly incorporated via a Markov random field model [8], [21], [34].…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

CollageParsing: Nonparametric Scene Parsing by Adaptive Overlapping Windows

Tung

Little

2014

Computer Vision – ECCV 2014

View full text Add to dashboard Cite

Abstract. Scene parsing is the problem of assigning a semantic label to every pixel in an image. Though an ambitious task, impressive advances have been made in recent years, in particular in scalable nonparametric techniques suitable for open-universe databases. This paper presents the CollageParsing algorithm for scalable nonparametric scene parsing. In contrast to common practice in recent nonparametric approaches, CollageParsing reasons about mid-level windows that are designed to capture entire objects, instead of low-level superpixels that tend to fragment objects. On a standard benchmark consisting of outdoor scenes from the LabelMe database, CollageParsing achieves state-of-the-art nonparametric scene parsing results with 7 to 11% higher average per-class accuracy than recent nonparametric approaches.

show abstract

Section: Related Workmentioning

confidence: 99%

Section: Forming the Retrieval Setmentioning

confidence: 99%

“…For θ we adopt the same pairwise potential term as in SuperParsing, which is based on the co-occurrences of category labels [34]: Fig. 4.…”

Section: Performing Mrf Inferencementioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

CollageParsing: Nonparametric Scene Parsing by Adaptive Overlapping Windows

Tung

Little

2014

Computer Vision – ECCV 2014

View full text Add to dashboard Cite

show abstract

CoralSeg: Learning coral segmentation from sparse annotations

Alonso

Yuval

Eyal

et al. 2019

Journal of Field Robotics

View full text Add to dashboard Cite

Robotic advances and developments in sensors and acquisition systems facilitate the collection of survey data in remote and challenging scenarios. Semantic segmentation, which attempts to provide per‐pixel semantic labels, is an essential task when processing such data. Recent advances in deep learning approaches have boosted this task's performance. Unfortunately, these methods need large amounts of labeled data, which is usually a challenge in many domains. In many environmental monitoring instances, such as the coral reef example studied here, data labeling demands expert knowledge and is costly. Therefore, many data sets often present scarce and sparse image annotations or remain untouched in image libraries. This study proposes and validates an effective approach for learning semantic segmentation models from sparsely labeled data. Based on augmenting sparse annotations with the proposed adaptive superpixel segmentation propagation, we obtain similar results as if training with dense annotations, significantly reducing the labeling effort. We perform an in‐depth analysis of our labeling augmentation method as well as of different neural network architectures and loss functions for semantic segmentation. We demonstrate the effectiveness of our approach on publicly available data sets of different real domains, with the emphasis on underwater scenarios—specifically, coral reef semantic segmentation. We release new labeled data as well as an encoder trained on half a million coral reef images, which is shown to facilitate the generalization to new coral scenarios.

show abstract

Effective Use of Synthetic Data for Urban Scene Semantic Segmentation

Saleh

Aliakbarian

Salzmann

et al. 2018

Lecture Notes in Computer Science

127

View full text Add to dashboard Cite

Training a deep network to perform semantic segmentation requires large amounts of labeled data. To alleviate the manual effort of annotating real images, researchers have investigated the use of synthetic data, which can be labeled automatically. Unfortunately, a network trained on synthetic data performs relatively poorly on real images. While this can be addressed by domain adaptation, existing methods all require having access to real images during training. In this paper, we introduce a drastically different way to handle synthetic images that does not require seeing any real images at training time. Our approach builds on the observation that foreground and background classes are not affected in the same manner by the domain shift, and thus should be treated differently. In particular, the former should be handled in a detection-based manner to better account for the fact that, while their texture in synthetic images is not photo-realistic, their shape looks natural. Our experiments evidence the effectiveness of our approach on Cityscapes and CamVid with models trained on synthetic data only.

show abstract

SuperParsing: Scalable Nonparametric Image Parsing with Superpixels

Cited by 429 publications

References 19 publications

CollageParsing: Nonparametric Scene Parsing by Adaptive Overlapping Windows

CollageParsing: Nonparametric Scene Parsing by Adaptive Overlapping Windows

CoralSeg: Learning coral segmentation from sparse annotations

Effective Use of Synthetic Data for Urban Scene Semantic Segmentation

Contact Info

Product

Resources

About