Peekaboo-Where are the Objects? Structure Adjusting Superpixels

Maierhofer, Georg; Heydecker, Daniel; Avilés-Rivero, Angelica I.; Alsaleli, Samar M.; Schönlieb, Carola‐Bibiane

doi:10.1109/icip.2018.8451822

Cited by 8 publications

(15 citation statements)

References 19 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Classic Techniques for Superpixels. Since the pioneering work of [28], several techniques have been proposed which can be roughly divided into patch-based models [7,33], watershed techniques [9,4,22], clustering-based approaches [1,17,21,2,23,44] and graph-based techniques [28,6,19,12]. The last two categories are the most widely applied family of techniques, which we will cover in the rest of this section.…”

Section: Related Workmentioning

confidence: 99%

“…For example, it unnecessarily partitions uniform areas and computes unnecessary distances in dense areas. These issues motivated several improvements for kmeans based techniques including reducing the number of distance calculations, improving the seeding initialisation and improving the feature representation [21,17,2,23,45].…”

Section: Related Workmentioning

confidence: 99%

“…The aforementioned advantages have encouraged the fast advancement of superpixel segmentation techniques (e.g. [16,1,19,17,23,36]) since the seminal work of Ren and Malik [28]. For a superpixel segmentation method to be useful, it should be computationally efficient and preserve the structure of the objects well (i.e.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

HERS Superpixels: Deep Affinity Learning for Hierarchical Entropy Rate Segmentation

Peng¹,

Avilés-Rivero²,

Schoenlieb

2022

2022 IEEE/CVF Winter Conference on Applications of Computer Vision (WACV)

Self Cite

View full text Add to dashboard Cite

Superpixels serve as a powerful preprocessing tool in many computer vision tasks. By using superpixel representation, the number of image primitives can be largely reduced by orders of magnitudes. The majority of superpixel methods use handcrafted features, which usually do not translate well into strong adherence to object boundaries. A few recent superpixel methods have introduced deep learning into the superpixel segmentation process. However, none of these methods is able to produce superpixels in near real-time, which is crucial to the applicability of a superpixel method in practice. In this work, we propose a two-stage graph-based framework for superpixel segmentation. In the first stage, we introduce an efficient Deep Affinity Learning (DAL) network that learns pairwise pixel affinities by aggregating multi-scale information. In the second stage, we propose a highly efficient superpixel method called Hierarchical Entropy Rate Segmentation (HERS). Using the learned affinities from the first stage, HERS builds a hierarchical tree structure that can produce any number of highly adaptive superpixels instantaneously. We demonstrate, through visual and numerical experiments, the effectiveness and efficiency of our method compared to various state-of-the-art superpixel methods.

show abstract

Section: Related Workmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

See 1 more Smart Citation

HERS Superpixels: Deep Affinity Learning for Hierarchical Entropy Rate Segmentation

Peng¹,

Avilés-Rivero²,

Schoenlieb

2022

2022 IEEE/CVF Winter Conference on Applications of Computer Vision (WACV)

Self Cite

View full text Add to dashboard Cite

show abstract

“…2) Hyperspectral clustering distance: Based on our previous work [39], we design a more effective clustering distance as a combination of the Euclidean spectral distance [34] and Log-Euclidean (LED) distance [40] of a covariance matrix representation [41]. This combination effectively combines the spatial and spectral data present in the image.…”

Section: A Superpixel Segmentationmentioning

confidence: 99%

Superpixel Contracted Graph-Based Learning for Hyperspectral Image Classification

Sellars

Avilés-Rivero

Schönlieb

2020

IEEE Trans. Geosci. Remote Sensing

Self Cite

View full text Add to dashboard Cite

A central problem in hyperspectral image classification is obtaining high classification accuracy when using a limited amount of labelled data. In this paper we present a novel graph-based framework, which aims to tackle this problem in the presence of large scale data input. Our approach utilises a novel superpixel method, specifically designed for hyperspectral data, to define meaningful local regions in an image, which with high probability share the same classification label. We then extract spectral and spatial features from these regions and use these to produce a contracted weighted graph-representation, where each node represents a region rather than a pixel. Our graph is then fed into a graph-based semi-supervised classifier which gives the final classification. We show that using superpixels in a graph representation is an effective tool for speeding up graphical classifiers applied to hyperspectral images. We demonstrate through exhaustive quantitative and qualitative results that our proposed method produces accurate classifications when an incredibly small amount of labelled data is used. We show that our approach mitigates the major drawbacks of existing approaches, resulting in our approach outperforming several comparative state-of-theart techniques.

show abstract

“…where (F (S i )) 1 represents that spatial part of the feature function, and | · | is the euclidean distance on the image grid. Furthermore, instead of using the Euclidean spectral distance found in [6,7], we instead use covariance matrix representation [8] and the Log-Euclidean distance (LED) [9] which is better suited for HSIs. For each pixel p ∈ I we construct a covariance matrix C p describing the relationship between different hyperspectral bands, which extracts powerful spectral and spatial information.…”

Section: Spectral Covariance Based Superpixelsmentioning

confidence: 99%

Semi-Supervised Learning with Graphs: Covariance Based Superpixels For Hyperspectral Image Classification

Sellars

Avilés-Rivero

Papadakis

et al. 2019

IGARSS 2019 - 2019 IEEE International Geoscience and Remote Sensing Symposium

Self Cite

View full text Add to dashboard Cite

In this paper, we present a graph-based semi-supervised framework for hyperspectral image classification. We first introduce a novel superpixel algorithm based on the spectral covariance matrix representation of pixels to provide a better representation of our data. We then construct a superpixel graph, based on carefully considered feature vectors, before performing classification. We demonstrate, through a set of experimental results using two benchmarking datasets, that our approach outperforms three state-of-the-art classification frameworks, especially when an extremely small amount of labelled data is used.

show abstract

Peekaboo-Where are the Objects? Structure Adjusting Superpixels

Cited by 8 publications

References 19 publications

HERS Superpixels: Deep Affinity Learning for Hierarchical Entropy Rate Segmentation

HERS Superpixels: Deep Affinity Learning for Hierarchical Entropy Rate Segmentation

Superpixel Contracted Graph-Based Learning for Hyperspectral Image Classification

Semi-Supervised Learning with Graphs: Covariance Based Superpixels For Hyperspectral Image Classification

Contact Info

Product

Resources

About