scholarly journals Can the Use of nonlinear Color Metrics systematically improve Segmentation?

2018 ◽  
Vol 25 (3) ◽  
pp. 23
Author(s):  
Luís Eduardo Ramos de Carvalho ◽  
Sylvio Luiz Mantelli Neto ◽  
Eros Comunello ◽  
Antonio Carlos Sobieranski ◽  
Aldo Von Wangenheim
Keyword(s):  
Mahalanobis Distance ◽  
Region Growing ◽  
Large Set ◽  
Distance Metrics ◽  
Distribution Centers ◽  
Color Structure ◽  
Systematic Improvement ◽  
L2 Norm ◽  
Segmentation Methods ◽  
Color Similarity

Image segmentation is a procedure where an image is split into its constituent parts, according to some criterion. In the literature, there are different well-known approaches for segmentation, such as clustering, thresholding, graph theory and region growing. Such approaches, additionally, can be combined with color distance metrics, playing an important role for color similarity computation. Aiming to investigate general approaches able to enhance the performance of segmentation methods, this work presents an empirical study of the effect of a nonlinear color metric on segmentation procedures. For this purpose, three algorithms were  chosen: Mumford-Shah, Color Structure Code and Felzenszwalb and Huttenlocher Segmentation. The color similarity metric employed by these algorithms (L2-norm) was replaced by the Polynomial Mahalanobis Distance. This metric is an extension of the statistical Mahalanobis Distance used to measure the distance between coordinates and distribution centers. An evaluation based upon automated comparison of segmentation results against ground truths from the Berkeley Dataset was performed. All three segmentation approaches were compared to their traditional implementations, against each other and also to a large set of other segmentation methods. The statistical analysis performed has indicated a systematic improvement of segmentation results for all three segmentation approaches when the nonlinear metric was employed.

scholarly journals Rooted Spanning Superpixels

2020 ◽  
Vol 128 (12) ◽  
pp. 2962-2978
Author(s):  
Dengfeng Chai
Keyword(s):  
Cost Function ◽  
State Of The Art ◽  
Region Growing ◽  
The Other ◽  
New Approach ◽  
Underlying Graph ◽  
Spanning Forest ◽  
Color Similarity ◽  
Path Cost ◽  
Number Of Seeds

Abstract This paper proposes a new approach for superpixel segmentation. It is formulated as finding a rooted spanning forest of a graph with respect to some roots and a path-cost function. The underlying graph represents an image, the roots serve as seeds for segmentation, each pixel is connected to one seed via a path, the path-cost function measures both the color similarity and spatial closeness between two pixels via a path, and each tree in the spanning forest represents one superpixel. Originating from the evenly distributed seeds, the superpixels are guided by a path-cost function to grow uniformly and adaptively, the pixel-by-pixel growing continues until they cover the whole image. The number of superpixels is controlled by the number of seeds. The connectivity is maintained by region growing. Good performances are assured by connecting each pixel to the similar seed, which are dominated by the path-cost function. It is evaluated by both the superpixel benchmark and supervoxel benchmark. Its performance is ranked as the second among top performing state-of-the-art methods. Moreover, it is much faster than the other superpixel and supervoxel methods.

scholarly journals Segmentation of cervical cell images.

1977 ◽  
Vol 25 (7) ◽  
pp. 681-688 ◽  
Author(s):  
R L Cahn ◽  
R S Poulsen ◽  
G Toussaint
Keyword(s):  
Mahalanobis Distance ◽  
Structural Information ◽  
Cervical Cytology ◽  
Contour Tracing ◽  
Cervical Cell ◽  
Segmentation Methods ◽  
Cervical Cells ◽  
The Stability ◽  
Cytology Screening ◽  
Segmentation Threshold

A major problem in the automation of cervical cytology screening is the segmentation of cell images. This paper describes various standard segmentation methods plus one which determines a segmentation threshold based on the stability of the perimeter of the cell as the threshold is varied. As well as contour, certain structural information is used to decide upon the threshold which separates cytoplasm from the background. Once the cytoplasm threshold is found, cytoplasm and nucleus are separated by simple clustering into three groups, cytoplasm, folded cytoplasm and nucleus. These techniques have been tested on 1500 cervical cells that belong to one of eight normal classes and five abnormal classes. A minimum Mahalanobis distance classifier was used to compare results. Manually thresholded cells were classified correctly 66.0% of the time for the 13 class problem and 95.2% of the time on the two (normal-abnormal) class problem. The contour tracing technique was 52.9% and 90.0% correct, respectively.

scholarly journals Wavelet adaptive quantization based color image segmentation

2020 ◽  
Author(s):  
Anand Swaminathan ◽  
K.Venkata Subramaniyan ◽  
Tiruppathirajan G. ◽  
Rajkumar J
Keyword(s):  
Image Segmentation ◽  
Color Image ◽  
Region Growing ◽  
Color Image Segmentation ◽  
Modern World ◽  
New Approach ◽  
Processing Step ◽  
Segmentation Algorithms ◽  
Color Similarity ◽  
Homogeneous Regions

Image segmentation is an important pre-processing step towards higher level tasks such as object recognition, computer vision or image compression. Most of the existing segmentation algorithms deal with grayscale images only. But in the modern world, color images are extensively used in many situations. A new approach for color image segmentation is presented in this paper. There are many ways to deal with image segmentation problem and in these techniques; a particular class of algorithms traces their origin from region-based methods. These algorithms group homogeneous pixels, which are connected to primitive regions. They are easy to implement and are promising. Therefore, here one of the most efficient region-based segmentation algorithms is explained. The color image is quantized adaptively, using a wavelet transform. Then the region growing process is adopted. As preprocess, before actual region merging, small regions are eliminated by merging them with neighbor regions depending upon color similarity. After this, homogeneous regions are merged to get segmented output.

scholarly journals Segmenting radiographs of the hand and wrist using computer vision

1990 ◽  
Author(s):  
Γεώργιος Μάνος
Keyword(s):  
Image Segmentation ◽  
Region Growing ◽  
Bone Age ◽  
Growth Disorders ◽  
Region Merging ◽  
Clinical Tool ◽  
Age Assessment ◽  
Segmentation Methods ◽  
Bone Age Assessment ◽  
Edge Based

"Bone age" age assessment is an important clinical tool in the area of paediatrics. The technique is based upon the appearance and growth of specific bones in a developing child. In particular most methods for "bone age" assessment are based on the examination of the growth of bones of the left hand and wrist on X-ray films. This assessment is useful in the treatment of growth disorders and also is used to predict adult height. One of the most reliable methods for "bone age" assessment is the TW2 method. The drawback of this method is that it is time consuming and therefore its automation is highly desirable. One of the most important aspects of the automation process is image segmentation i.e. the extraction of bones from soft-tissue and background. Over the past 10 years various attempts have been made at the segmentation of handwrist radiographs but with limited success. This can mainly be attributed to the characteristics of the scenes e.g. biological objects, penetrating nature of radiation, faint bone boundaries, uncertainty of scene content, and conjugation of bones. Experience in the field of radiographic image analysis has shown thatsegmentation of radiographic scenes is a difficult task requiring solutions which depend on the nature of the particular problem.There are two main approaches to image segmentation: edge based and region based. Most of the previous attempts at the segmentation of hand-wrist radiographs were edge based. Edge based methods usually require a w-ell defined model of the object boundaries in order to produce successful results. However, for this particular application it is difficult to derive such a model. Region based segmentation methods have produced promising results for scenes which exhibit uncertainty regarding their content and boundaries of objects in the image, as in the case, for example, of natural senes. This thesis presents a segmentation method based on the concept of regions. This method consists of region growing and region merging stages. A technique was developed for region merging which combines edge and region boundai^ information. A bone extraction stage follows which labels regions as either boneor background using heuristic rules based on the grey-level properties of the scene. Finally, a technique is proposed for the segmentation of bone outlines which helps in identifying conjugated bones. Experimental results have demonstrated that this method represents a significant improvement over existing segmentation methods for hand-wrist radiographs, particularly with regard to the segmentation of radiographs with varying degrees of bone maturity.

scholarly journals Segmentation of Left and Right Ventricles in Cardiac MRI Using Active Contours

2017 ◽  
Vol 2017 ◽  
pp. 1-16 ◽  
Author(s):  
Shafiullah Soomro ◽  
Farhan Akram ◽  
Asad Munir ◽  
Chang Ha Lee ◽  
Kwang Nam Choi
Keyword(s):  
Cardiac Mri ◽  
Active Contours ◽  
Validation Dataset ◽  
Distance Metrics ◽  
Pressure Force ◽  
Segmentation Methods ◽  
Quantitative Validation ◽  
Left And Right ◽  
Magnetic Resonance Imaging Mri ◽  
Dice Index

Segmentation of left and right ventricles plays a crucial role in quantitatively analyzing the global and regional information in the cardiac magnetic resonance imaging (MRI). In MRI, the intensity inhomogeneity and weak or blurred object boundaries are the problems, which makes it difficult for the intensity-based segmentation methods to properly delineate the regions of interests (ROI). In this paper, a hybrid signed pressure force function (SPF) is proposed, which yields both local and global image fitted differences in an additive fashion. A characteristic term is also introduced in the SPF function to restrict the contour within the ROI. The overlapping dice index and Hausdorff-Distance metrics have been used over cardiac datasets for quantitative validation. Using 2009 LV MICCAI validation dataset, the proposed method yields DSC values of 0.95 and 0.97 for endocardial and epicardial contours, respectively. Using 2012 RV MICCAI dataset, for the endocardial region, the proposed method yields DSC values of 0.97 and 0.90 and HD values of 8.51 and 7.67 for ED and ES, respectively. For the epicardial region, it yields DSC values of 0.92 and 0.91 and HD values of 6.47 and 9.34 for ED and ES, respectively. Results show its robustness in the segmentation application of the cardiac MRI.

Improving Graph-Based Image Segmentation Using Nonlinear Color Similarity Metrics

2015 ◽  
Vol 15 (04) ◽  
pp. 1550018 ◽  
Author(s):  
L. E. Carvalho ◽  
S. L. Mantelli Neto ◽  
A. C. Sobieranski ◽  
E. Comunello ◽  
A. von Wangenheim
Keyword(s):  
Similarity Metrics ◽  
Empirical Validation ◽  
Discrimination Function ◽  
Segmentation Method ◽  
Dissimilarity Index ◽  
Segmentation Methods ◽  
Segmentation Quality ◽  
Color Similarity ◽  
Golden Standard ◽  
Validation Experiments

We present a new segmentation method called weighted Felzenszwalb and Huttenlocher (WFH), an improved version of the well-known graph-based segmentation method, Felzenszwalb and Huttenlocher (FH). Our algorithm uses a nonlinear discrimination function based on polynomial Mahalanobis Distance (PMD) as the color similarity metric. Two empirical validation experiments were performed using as a golden standard ground truths (GTs) from a publicly available source, the Berkeley dataset, and an objective segmentation quality measure, the Rand dissimilarity index. In the first experiment the results were compared against the original FH method. In the second, WFH was compared against several well-known segmentation methods. In both cases, WFH presented significant better similarity results when compared with the golden standard and segmentation results presented a reduction of over-segmented regions.

Sign in / Sign up

Export Citation Format

Share Document