Fast marching based superpixels

2020 ◽  
Vol 4 (1) ◽  
pp. 127-142
Author(s):  
Kaiwen Chang ◽  
Bruno Figliuzzi
Keyword(s):  
Heterogeneous Medium ◽  
Region Growing ◽  
Fast Marching ◽  
Local Color ◽  
Superpixel Segmentation ◽  
Spanning Forest ◽  
Texture Information ◽  
Segmentation Algorithms ◽  
Simple Linear Iterative Clustering ◽  
Better Than

AbstractIn this article, we present a fast-marching based algorithm for generating superpixel (FMS) partitions of images. The idea behind the algorithm is to draw an analogy between waves propagating in a heterogeneous medium and regions growing on an image at a rate depending on the local color and texture. The FMS algorithm is evaluated on the Berkeley Segmentation Dataset 500. It yields results in terms of boundary adherence that are slightly better than the ones obtained with similar approaches including the Simple Linear Iterative Clustering, the Eikonal-based region growing for efficient clustering and the Iterative Spanning Forest framework for superpixel segmentation algorithms. An interesting feature of the proposed algorithm is that it can take into account texture information to compute the superpixel partition. We illustrate the interest of adding texture information on a specific set of images obtained by recombining textures patches extracted from images representing stripes, originally constructed by Giraud et al. [20]. On this dataset, our approach works significantly better than color based superpixel algorithms.

scholarly journals GRID: GRID Resample by Information Distribution

Symmetry ◽  
2020 ◽  
Vol 12 (9) ◽  
pp. 1417
Author(s):  
Cheng Li ◽  
Baolong Guo ◽  
Zhe Huang ◽  
Jianglei Gong ◽  
Xiaodong Han ◽  
...  
Keyword(s):  
Local Concentration ◽  
Context Aware ◽  
Color Information ◽  
Expected Number ◽  
Local Color ◽  
Grid Sampling ◽  
Color Uniformity ◽  
Segmentation Algorithms ◽  
Simple Linear Iterative Clustering ◽  
Coarse To Fine

This paper exploits a concise yet efficient initialization strategy to optimize grid sampling-based superpixel segmentation algorithms. Rather than straight distributing all initial seeds evenly, it adopts a context-aware approach to modify their positions and total number via a coarse-to-fine manner. Firstly, half the expected number of seeds are regularly sampled on the image grid, thereby creating a rough distribution of color information for all rectangular cells. A series of fission is then performed on cells that contain excessive color information recursively. In each cell, the local color uniformity is balanced by a dichotomy on one original seed, which generates two new seeds and settles them to spatially symmetrical sub-regions. Therefore, the local concentration of seeds is adaptive to the complexity of regional information. In addition, by calculating the amount of color via a summed area table (SAT), the informative regions can be located at a very low time cost. As a result, superpixels are produced from ideal original seeds with an exact number and exhibit better boundary adherence. Experiments demonstrate that the proposed strategy effectively promotes the performance of simple linear iterative clustering (SLIC) and its variants in terms of several quality measures.

scholarly journals CONIC: Contour Optimized Non-Iterative Clustering Superpixel Segmentation

2021 ◽  
Vol 13 (6) ◽  
pp. 1061
Author(s):  
Cheng Li ◽  
Baolong Guo ◽  
Nannan Liao ◽  
Jianglei Gong ◽  
Xiaodong Han ◽  
...  
Keyword(s):  
State Of The Art ◽  
Visual Quality ◽  
Local Optimum ◽  
Trade Off ◽  
Grid Sampling ◽  
Superpixel Segmentation ◽  
Segmentation Accuracy ◽  
Segmentation Algorithms ◽  
Color Similarity ◽  
Better Than

Superpixels group perceptually similar pixels into homogeneous sub-regions that act as meaningful features for advanced tasks. However, there is still a contradiction between color homogeneity and shape regularity in existing algorithms, which hinders their performance in further processing. In this work, a novel Contour Optimized Non-Iterative Clustering (CONIC) method is presented. It incorporates contour prior into the non-iterative clustering framework, aiming to provide a balanced trade-off between segmentation accuracy and visual uniformity. After the conventional grid sampling initialization, a regional inter-seed correlation is first established by the joint color-spatial-contour distance. It then guides a global redistribution of all seeds to modify the number and positions iteratively. This is done to avoid clustering falling into the local optimum and achieve the exact number of user-expectation. During the clustering process, an improved feature distance is elaborated to measure the color similarity that considers contour constraint and prevents the boundary pixels from being wrongly assigned. Consequently, superpixels acquire better visual quality and their boundaries are more consistent with the object contours. Experimental results show that CONIC performs as well as or even better than the state-of-the-art superpixel segmentation algorithms, in terms of both efficiency and segmentation effects.

scholarly journals Liver tumour classification using average correction higher order local autocorrelation coefficient and legendre moments

2018 ◽  
Vol 7 (2.6) ◽  
pp. 306
Author(s):  
Aravinda H.L ◽  
M.V Sudhamani
Keyword(s):  
Texture Feature ◽  
Region Growing ◽  
Ct Images ◽  
Higher Order ◽  
Liver Carcinoma ◽  
Tumor Segmentation ◽  
Abdominal Ct ◽  
Simple Linear Iterative Clustering ◽  
Classification Of Tumors ◽  
Legendre Moments

The major reasons for liver carcinoma are cirrhosis and hepatitis.  In order to  identify carcinoma in the liver abdominal CT images are used. From abdominal CT images, segmentation of liver portion using adaptive region growing, tumor segmentation from extracted liver using Simple Linear Iterative Clustering is already implemented. In this paper, classification of tumors as benign or malignant is accomplished using Rough-set classifier based on texture feature extracted using Average Correction Higher Order Local Autocorrelation Coefficients and Legendre moments. Classification accuracy achieved in proposed scheme is 90%. The results obtained are promising and have been compared with existing methods.

Content-Based Video Retrieval Using Integration of Curvelet Transform and Simple Linear Iterative Clustering

2021 ◽  
pp. 2250018
Author(s):  
Reddy Mounika Bommisetty ◽  
Ashish Khare ◽  
Manish Khare ◽  
P. Palanisamy
Keyword(s):  
Video Retrieval ◽  
Information Source ◽  
Pearson Correlation ◽  
Curvelet Transform ◽  
Video On Demand ◽  
Traffic Monitoring ◽  
Superpixel Segmentation ◽  
Online Processing ◽  
Simple Linear Iterative Clustering ◽  
Content Based Video Retrieval

Video is a rich information source containing both audio and visual information along with motion information embedded in it. Applications such as e-learning, live TV, video on demand, traffic monitoring, etc. need an efficient video retrieval strategy. Content-based video retrieval and superpixel segmentation are two diverse application areas of computer vision. In this work, we are presenting an algorithm for content-based video retrieval with help of Integration of Curvelet transform and Simple Linear Iterative Clustering (ICTSLIC) algorithm. Proposed algorithm consists of two steps: off line processing and online processing. In offline processing, keyframes of the database videos are extracted by employing features: Pearson Correlation Coefficient (PCC) and color moments (CM) and on the extracted keyframes superpixel generation algorithm ICTSLIC is applied. The superpixels generated by applying ICTSLIC on keyframes are used to represent database videos. On other side, in online processing, ICTSLIC superpixel segmentation is applied on query frame and the superpixels generated by segmentation are used to represent query frame. Then videos similar to query frame are retrieved through matching done by calculation of Euclidean distance between superpixels of query frame and database keyframes. Results of the proposed method are irrespective of query frame features such as camera motion, object’s pose, orientation and motion due to the incorporation of ICTSLIC superpixels as base feature for matching and retrieval purpose. The proposed method is tested on the dataset comprising of different categories of video clips such as animations, serials, personal interviews, news, movies and songs which is publicly available. For evaluation, the proposed method randomly picks frames from database videos, instead of selecting keyframes as query frames. Experiments were conducted on the developed dataset and the performance is assessed with different parameters Precision, Recall, Jaccard Index, Accuracy and Specificity. The experimental results shown that the proposed method is performing better than the other state-of-art methods.

scholarly journals Superpixel Segmentation Using Dynamic and Iterative Spanning Forest

2020 ◽  
Vol 27 ◽  
pp. 1440-1444
Author(s):  
Felipe C. Belem ◽  
Silvio Jamil F. Guimaraes ◽  
Alexandre X. Falcao
Keyword(s):  
Superpixel Segmentation ◽  
Spanning Forest

PERFORMANCE EVALUATION OF REGION-GROWING BASED SEGMENTATION ALGORITHMS FOR SEGMENTING THE AORTA

2016 ◽  
Vol 78 (4-3) ◽  
Author(s):  
Hussain Rahman ◽  
Fakhrud Din ◽  
Sami ur Rahmana ◽  
Sehatullah Sehatullah
Keyword(s):  
Image Segmentation ◽  
Medical Images ◽  
Region Growing ◽  
Medical Image Segmentation ◽  
Ease Of Use ◽  
Daily Routine ◽  
Segmentation Algorithms ◽  
Abdominal Mri ◽  
Mri Scans ◽  
The Common

Region-growing based image segmentation techniques, available for medical images, are reviewed in this paper. In digital image processing, segmentation of humans' organs from medical images is a very challenging task. A number of medical image segmentation techniques have been proposed, but there is no standard automatic algorithm that can generally be used to segment a real 3D image obtained in daily routine by the clinicians. Our criteria for the evaluation of different region-growing based segmentation algorithms are: ease of use, noise vulnerability, effectiveness, need of manual initialization, efficiency, computational complexity, need of training, information used, and noise vulnerability. We test the common region-growing algorithms on a set of abdominal MRI scans for the aorta segmentation. The evaluation results of the segmentation algorithms show that region-growing techniques can be a better choice for segmenting an object of interest from medical images.

scholarly journals A Hybrid Method for Pancreas Extraction from CT Image Based on Level Set Methods

2013 ◽  
Vol 2013 ◽  
pp. 1-13 ◽  
Author(s):  
Huiyan Jiang ◽  
Hanqing Tan ◽  
Hiroshi Fujita
Keyword(s):  
Level Set Method ◽  
Level Set ◽  
Negative Impact ◽  
Level Set Methods ◽  
Region Growing ◽  
Ct Images ◽  
Medical Image Segmentation ◽  
Initial Contour ◽  
Ct Image ◽  
Fast Marching

This paper proposes a novel semiautomatic method to extract the pancreas from abdominal CT images. Traditional level set and region growing methods that request locating initial contour near the final boundary of object have problem of leakage to nearby tissues of pancreas region. The proposed method consists of a customized fast-marching level set method which generates an optimal initial pancreas region to solve the problem that the level set method is sensitive to the initial contour location and a modified distance regularized level set method which extracts accurate pancreas. The novelty in our method is the proper selection and combination of level set methods, furthermore an energy-decrement algorithm and an energy-tune algorithm are proposed to reduce the negative impact of bonding force caused by connected tissue whose intensity is similar with pancreas. As a result, our method overcomes the shortages of oversegmentation at weak boundary and can accurately extract pancreas from CT images. The proposed method is compared to other five state-of-the-art medical image segmentation methods based on a CT image dataset which contains abdominal images from 10 patients. The evaluated results demonstrate that our method outperforms other methods by achieving higher accuracy and making less false segmentation in pancreas extraction.

The Technique of Shape-Based Multi-Feature Combination of TradeMark Image Retrieval

2012 ◽  
Vol 429 ◽  
pp. 287-291 ◽  
Author(s):  
Cong Zhang ◽  
Fu Cheng You
Keyword(s):  
Image Retrieval ◽  
Euclidean Distance ◽  
Feature Matching ◽  
Region Growing ◽  
Local Information ◽  
Feature Combination ◽  
Single Feature ◽  
Hu Moments ◽  
Global Matching ◽  
Better Than

At present, the technique of trademark image retrieval based on multi-feature combination of the shape mainly includes single-feature global matching or local matching and multi-feature matching, which is playing a more and more important role in the area of the trademark image retrieval. In this paper, due to the deficiency described by some single shape-based features, the technique of the multi-feature combination trademark image retrieval is proposed based on the region and the edge of a shape. Firstly, a trademark image is segmented with region growing, then low order Hu moments and eccentricity are extracted on the resulting region, which is able to express the local information of the image; Secondly, there is an extraction of Compactness and Convexity, which describe the global feature of the image, on the edge extracted with Canny. At last, the combination of the multi-feature is applied to get a Euclidean distance. Good results have been obtained in the following experiment, which proves the multi-feature combination way is better than other single-feature ways.

Research on Threshold Segmentation Algorithms

2013 ◽  
Vol 860-863 ◽  
pp. 2888-2891
Author(s):  
Yu Bing Dong ◽  
Ming Jing Li ◽  
Ying Sun
Keyword(s):  
Pattern Recognition ◽  
Image Segmentation ◽  
Segmentation Method ◽  
Threshold Segmentation ◽  
Segmentation Methods ◽  
Segmentation Algorithms ◽  
Better Than ◽  
Iterative Threshold

Thresholding is one of the critical steps in pattern recognition and has a significant effect on the upcoming steps of image application, the important objectives of thresholding are as follows, and separating objects from background, decreasing the capacity of data consequently increases speed. Various threshold segmentation methods are studied. These methods are compared by using MATLAB7.0. The qualities of image segmentation are elaborated. The results show that iterative threshold segmentation method is better than others.

Sign in / Sign up

Export Citation Format

Share Document