Content-sensitive superpixel segmentation via self-organization-map neural network☆
Introduction
A superpixel is defined as a set of grouped homogeneous pixels in an image. Superpixel segmentation is also called as oversegmentation. It is different from image segmentation [1], [2] and co-segmentation [3]. Both image segmentation and co-segmentation intend to extract foreground from images, but co-segmentation always segments a series images with similar scene simultaneously. However, superpixel segmentation aims to group pixels into homogeneous regions for the following tasks. Superpixels are used to replace pixels as the atomic unit in the following computer vision tasks. Some image segmentation algorithms can be used to extract superpixels from image, such as QuickShift [4], MeanShift [5] and Normalized cuts [6]. Since it can be used to improve the performance of subsequent computer vision tasks, superpixel segmentation has attracted more and more attentions in recent years and has been widely applied to many applications, such as saliency detection [7], [8], image segmentation [9], [10], image parsing [11], surface reconstruction [12] and object recognition [13], [14].
Many superpixel segmentation methods have been proposed [15], [16], [17], [18], [19], [20], [21], [22], [23], [24]. However, the following challenge of superpixel segmentation is still unsolved: on one hand superpixel segmentation should avoid under-segmentation error and preserve the details of image content; on the other hand, superpixel segmentation requires generating superpixels as few as possible to reduce the computation complexity of the following tasks. This motivates researchers to focus on content-sensitive superpixels segmentation [19], [23].
Content-sensitive superpixel whose size is sensitive to the content density of local regions in the image. The content density of an image region is usually measured by its color variation. It often differs in different parts of the image. Fig. 1 shows an example of such situation. The color variation of pixels in the green window in Fig. 1 is smaller than that in the red window. Traditional superpixel segmentation algorithms ignore such difference. The superpixels they generated usually have similar size and shape. On the other hand, content-sensitive approaches can produce small superpixels in regions have high color variation and large superpixels in regions have low color variation. As pointed out by Liu [19], this is a better image representation than traditional uniform superpixel, which has lower under-segmentation error and can preserve more details of image boundary.
The previous methods of content-sensitive superpixel segmentation include SSS [23] and Manifold SLIC [19]. In the SSS, two steps are used to generate superpixels. First, the centers of each superpixel are roughly placed in a lattice structure on the image. Then, the centers are relocated or split repeatedly until the energy function which is defined on geodesic distances between pixels and the compactness of superpixels meets the termination condition. The energy function consists of two terms. The first term embeds the color homogeneity of pixels and the second item integrates the content density and compactness constraints. The content-sensitiveness of superpixel can be adjusted using the balance factor between the two terms. Manifold SLIC extends traditional superpixel algorithm of Simple Linear Iterative Clustering (SLIC) to generate content-sensitive superpixels by mapping the image into a 2-dimensional manifold, in which the content density of the image is measured. Actually, it uses an efficient algorithm to compute restricted centroidal Voronoi tessellation on the manifold. Then the content-sensitiveness of superpixels is measured by the areas of Voronoi cells on the 2-dimensional manifold after mapping and the content-sensitive superpixels are produced accordingly. Manifold SLIC runs 10 times faster than SSS. Although the two methods above try to compute the content-sensitive superpixels, the explicit measurement of content-sensitiveness has not been explored. In this paper, we introduce a metric to measure the content-sensitiveness of superpixels, based on which we propose a superpixel segmentation method by using Self-Organization Map (SOM) neural network. We call the proposed method SOMS (SOM Superpixel) for short. Actually, we develop a content-sensitive sampling method to get pixels from the image. Then these pixels are used to train a SOM neural network for clustering pixels into content-sensitive superpixels. The main contributions of this paper are summarized as follows:
- 1.
A novel metric to measure the content-sensitiveness of superpixels is proposed.
- 2.
With this novel metric, we put forward a content-sensitive sampling algorithm to get pixels from images.
- 3.
We present the SOMS algorithm by training a SOM on the sampled pixels and use it to cluster image into superpixels.
This paper is organized as follows. In Section 2, we describe our SOMS algorithm. In Section 3, our approach is thoroughly evaluated and compared with other state-of-the-art approaches. Section 4 concludes this paper.
Section snippets
The proposed method
Superpixel segmentation algorithms can be roughly divided into two categories: graph based and clustering based. The clustering-based approach groups pixels into clusters and iteratively refine them to get superpixels. The main clustering methods employed in pervious superpixel segmentation include k-means, spectral clustering, and DBSCAN. In this paper, we resort to Self-Organized Map [25] for completing the clustering. The steps of SOMS algorithm are shown in Fig. 2 and described as follows:
- •
Datasets
We conduct the experiments on the Berkeley Segmentation Dataset (BSD) [28] and INRIA dataset [29]. We use Berkeley Segmentation Dataset to evaluate the accuracy of superpixel segmentation and use INRIA dataset to evaluate the efficiency of these algorithms for high-resolution images. BSD is a popular dataset for image segmentation evaluation, which is widely used to evaluate the accuracy of superpixel segmentation [15], [20], [24]. BSD dataset is consisted of 500 natural images. It is split
Conclusions
In this paper, we have proposed a content-sensitive superpixel segmentation algorithm based on Self-Origination Map (SOM), called SOMS for short. We introduce a novel metric to evaluate the content-sensitiveness of superpixels. Based on this metric, we present a content-sensitive sampling algorithm to sample pixels from the image and use them to train SOM neural network. Then, the trained SOM is used to cluster the pixels in the image into superpixels. The evaluation was conducted on Berkeley
Declaration of Competing Interest
The authors declared that there is no conflict of interest.
References (34)
- et al.
An efficient self-organizing active contour model for image segmentation
Neurocomputing
(2015) Self-organizing neural projections
Neural Netw.
(2006)Self-organizing map algorithm and distortion measure
Neural Netw.
(2006)Content-based image collection summarization and comparison using self-organizing maps
Pattern Recogn.
(2007)- et al.
Superpixel segmentation: a benchmark
Signal Process.: Image Commun.
(2017) - et al.
Segnet: a deep convolutional encoder-decoder architecture for image segmentation
IEEE Trans. Pattern Anal. Mach. Intell.
(2017) - et al.
Unsupervised multi-class joint image segmentation
- et al.
Quick shift and kernel methods for mode seeking
- et al.
Mean shift: a robust approach toward feature space analysis
IEEE Trans. Pattern Anal. Mach. Intell.
(2002) - et al.
Normalized cuts and image segmentation
IEEE Trans. Pattern Anal. Mach. Intell.
(2000)
Saliencyrank: two-stage manifold ranking for salient object detection
Comput. Vis. Media
Co-saliency detection via partially absorbing random walk
Co-segmentation of aircrafts from high-resolution satellite images
Object co-segmentation via graph optimized-flexible manifold ranking
Superparsing: scalable nonparametric image parsing with superpixels
Superpixel meshes for fast edge-preserving surface reconstruction
Maximum cohesive grid of superpixels for fast object localization
Cited by (6)
Multiscale superpixel method for segmentation of breast ultrasound
2020, Computers in Biology and MedicineCitation Excerpt :The reason for decomposition into 40 superpixels was because combining the decomposition with the distance transform adheres well to small and large tumours. Furthermore, from the structure of the superpixels, when m is approximately (10–40), the resulting superpixels adhere more to image boundaries [22]; hence, the decomposition to 40 superpixels adheres well to both small and large boundaries. Three radiologists delineated the ground truths using a Samsung Galaxy tablet and the electronic pen.
FuSC: Fusing Superpixels for Improved Semantic Consistency
2024, IEEE AccessAnalysis of dual-stage filtration and validation of high-dimensional real process data for creation of machine learning algorithms
2021, International Conference on Electrical, Computer, Communications and Mechatronics Engineering, ICECCME 2021
- ☆
This paper has been recommended for acceptance by Zicheng Liu.
- 1
This work was supported in part by National Natural Science Foundation of China [grant numbers 60973059, 81171407] and Program for New Century Excellent Talents in University of China [grant number NCET-10-0044].
- 2
Principal corresponding author.