Spatial adjacent bag of features with multiple superpixels for object segmentation and classification

Tao, Wenbing; Zhou, Yicong; Liu, Liman; Li, Kunqian; Sun, Kun; Zhang, Zhiguo

doi:10.1016/j.ins.2014.05.032

Cited by 13 publications

(5 citation statements)

References 48 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Superpixels have been used as features for classification, segmentation, and tracking tasks. In image classification or segmentation, they have been found to be useful for remote sensing [1], object classification [2], motion words for videos of hyperspectral images [3], medical imaging [4] and medical segmentation [5]. In the era of video analysis, segmentation has dominated superpixel-based feature approaches, such as video segmentation using superpixel flows [6], learning to segment moving objects [7], perceptual organization [8], and temporal superpixels using tracking figure-ground segments [9].…”

Section: Related Workmentioning

confidence: 99%

Motion Superpixels for Temporal Video Classification

Yudistira¹,

Kurita²

2018

Preprint

View full text Add to dashboard Cite

Superpixels are a representation of still images as pixel grids because of their more meaningful information compared with atomic pixels. However, their usefulness for video classification has been given little attention. In this paper, rather than using spatial RGB values as low-level features, we use optical flows mapped into hue-saturation-value (HSV) space to capture rich motion features over time. We introduce motion superpixels, which are superpixels generated from flow fields. After mapping flow fields into HSV space, independent superpixels are formed by iteration of seeded regions. Every grid of a motion superpixel is tracked over time using nearest neighbors in the histogram of flow (HOF) for consecutive flow fields. To define the temporal representation, the evolution of three features within the superpixel region, namely the HOF, HOG, and the center of superpixel mass are used as descriptors. The bag of features algorithm is used to quantify final features, and generalized histogram-kernel support vector machines are used as learning algorithms. We evaluate the proposed superpixel tracking on first-person videos and action sports videos.

show abstract

Section: Related Workmentioning

confidence: 99%

Motion Superpixels for Temporal Video Classification

Yudistira¹,

Kurita²

2018

Preprint

View full text Add to dashboard Cite

show abstract

“…Recently, higher-order information has been widely utilized for image segmentation [36,37,39,42,44,60] and image classification [54,66,67,68]. Compared with the pixel-level information, the higher-order information has the capability to model complex interactions of random variables [42], which leads to more informative results.…”

Section: Introductionmentioning

confidence: 99%

Label propagation and higher-order constraint-based segmentation of fluid-associated regions in retinal SD-OCT images

Wang

Sun

et al. 2016

Information Sciences

View full text Add to dashboard Cite

The segmentation of the fluid-associated region in the retina plays an important role in the treatment of retinal diseases, such as diabetic retinopathy (DR) and age-related macular degeneration (AMD). The existing methods for the detection of the fluid region generally suffer from the limitation of the segmentation accuracy and the expensive time costs. To overcome these problems, in this paper, we propose an interactive segmentation method for the fluid-associated region in three-dimensional (3-D) spectral domain optical coherence tomography (SD-OCT) retinal imaging, in which only a few seeds in one SD-OCT slice are needed, after which the algorithm finishes the segmentation automatically. To improve the segmentation accuracy, the higher-order constraint is introduced into the conventional Markov random field (MRF) framework to impose the superpixel consistency. To maintain temporal coherence of the 3-D SD-OCT slices, the labeling information is propagated slice by slice via a proposed motion-estimation-based algorithm. The proposed higher-order-based energy function can be efficiently solved by the max flow algorithm on a specified graph with several auxiliary nodes. Experiments on 28 SD-OCT cubes demonstrate the competitiveness of the proposed method compared with the state-of-the-art methods.

show abstract

“…Even though these low-level features have shown favorable results, they are unbefitting in many complex scenes, such as texture, luminance, and overlapping. To deal with this problem, some researchers attempted to combine different features together [9]. However, most of them integrated these features in a simple way.…”

Section: Introductionmentioning

confidence: 99%

Objectness Supervised Merging Algorithm for Color Image Segmentation

Sima

Wang

et al. 2016

Mathematical Problems in Engineering

View full text Add to dashboard Cite

Ideal color image segmentation needs both low-level cues and high-level semantic features. This paper proposes a two-hierarchy segmentation model based on merging homogeneous superpixels. First, a region growing strategy is designed for producing homogenous and compact superpixels in different partitions. Total variation smoothing features are adopted in the growing procedure for locating real boundaries. Before merging, we define a combined color-texture histogram feature for superpixels description and, meanwhile, a novel objectness feature is proposed to supervise the region merging procedure for reliable segmentation. Both color-texture histograms and objectness are computed to measure regional similarities between region pairs, and the mixed standard deviation of the union features is exploited to make stop criteria for merging process. Experimental results on the popular benchmark dataset demonstrate the better segmentation performance of the proposed model compared to other well-known segmentation algorithms.

show abstract

Spatial adjacent bag of features with multiple superpixels for object segmentation and classification

Cited by 13 publications

References 48 publications

Motion Superpixels for Temporal Video Classification

Motion Superpixels for Temporal Video Classification

Label propagation and higher-order constraint-based segmentation of fluid-associated regions in retinal SD-OCT images

Objectness Supervised Merging Algorithm for Color Image Segmentation

Contact Info

Product

Resources

About