Fast K-Means Color Image Clustering With Normalized Distance Values

Baykan, Nurdan Akhan; Sağlam, Ali

doi:10.15317/scitech.2018.124

Cited by 3 publications

(6 citation statements)

References 13 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…NDVs are taken from different attribute values (also transformed into different color spaces) with different distance norms. The histogram-based k-means algorithm has been applied to NDVs with the upper limit iteration number 50 [4]). The number of clusters has been selected by considering the number of classes in GT images.…”

Section: Clustering Resultsmentioning

confidence: 99%

“…This method has been proposed to process the multidimensional data on a histogram because applying some methods (e.g., the k-means clustering) on very large amounts of data with less processing needs a histogram-based approach. A histogram indicates the frequencies of the data elements having the same attribute values [3,4]. The attribute values in a histogram must be normalized in a specified range (the closed range 0 − 255 is used in this study).…”

Section: Normalized Distance Values (Ndvs)mentioning

confidence: 99%

“…The NDVs of a point in a multidimensional space is the normalized and rounded distance from the point to the center of all points in a specified space [4]. NDVs are used to reduce the vector values that consist of more than one coordinate values into one value.…”

Section: Normalized Distance Values (Ndvs)mentioning

confidence: 99%

“…This process accelerates the clustering algorithm by preventing proximity calculation for all elements. To express the data with a useful histogram, each data element must have a single property value and the value must be discrete in a specified range; for example the integers in the 8-bit pixel value interval range from 0 to 255 [4].…”

Section: Introductionmentioning

confidence: 99%

“…Some of the simple ways to express these pixels with histogram are the procedure of converting the color value vectors to the gray values or the procedure of using one attribute band. Baykan and Saglam proposed a histogram-based method that makes use of all attribute values of image pixels and applied it on remote sensing images [4]. The results show that the method gives more accurate clustering results compared to the gray-level histogram.…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

Evaluating the attributes of remote sensing image pixels for fast k-meansclustering

Sağlam¹,

Baykan²

2019

Turk J Elec Eng & Comp Sci

Self Cite

View full text Add to dashboard Cite

Clustering process is an important stage for many data mining applications. In this process, data elements are grouped according to their similarities. One of the most known clustering algorithms is the k-means algorithm.The algorithm initially requires the number of clusters as a parameter and runs iteratively. Many remote sensing image processing applications usually need the clustering stage like many image processing applications. Remote sensing images provide more information about the environments with the development of the multispectral sensor and laser technologies. In the dataset used in this paper, the infrared (IR) and the digital surface maps (DSM) are also supplied besides the red (R), the green (G), and the blue (B) color values of the pixels. However, remote sensing images come with very large sizes (6000 × 6000 pixels for each image in the dataset used). Clustering these large-size images using their multiattributes consumes too much time if it is used directly. In the literature, some studies are available to accelerate the k-means algorithm. One of them is the normalized distance value (NDV)-based fast k-means algorithm that benefits from the speed of the histogram-based approach and uses the multiattributes of the pixels. In this paper, we evaluated the effects of these attributes on the correctness of the clustering process with different color space transformations and distance measurements. We give the success results as peak signal-to-noise ratio and structural similarity index values using two different types of reference data (the source images and the ground-truth images) separately. Finally, we give the results based on accuracy measurement for evaluating both the success of the clustering outputs and the reliability of the NDV-based measurement methods presented in this paper.

show abstract

Section: Clustering Resultsmentioning

confidence: 99%

Section: Normalized Distance Values (Ndvs)mentioning

confidence: 99%

Section: Normalized Distance Values (Ndvs)mentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Evaluating the attributes of remote sensing image pixels for fast k-meansclustering

Sağlam¹,

Baykan²

2019

Turk J Elec Eng & Comp Sci

Self Cite

View full text Add to dashboard Cite

show abstract

Roof detection on satellite images

Sağlam

Baykan

2017

2017 25th Signal Processing and Communications Applications Conference (SIU)

View full text Add to dashboard Cite

A novel Histogram Image Clustering Approach using Enhanced Firefly Algorithm with K-means and expanded exploitation of Aquila Optimizer

Dhal,

Das,

Galvez

et al. 2024

Preprint

View full text Add to dashboard Cite

One of the most popular methods used in the field of image segmentation is K-means (KM). However, some limitations are presented, the computational time and the initialization process of the cluster centers. This study provides a Histogram-Based KM (HBKM) clustering approach that incorporates a modified Firefly Algorithm (FA) to overcome the KM drawbacks. In the histogram-based method, it is implemented considering grey-level histograms rather than image pixels. As a result, time complexity significantly decreases due to the number of grey levels employed. Moreover, the original KM initialization procedure is prone to be trapped in local optima. Consequently, the proposed approach can avoid this issue based on the exploitation and exploration mechanisms of the Aquila Optimizer (AO) method. A rigorous experimental analysis for comparing the performance of the proposed method against several state-of-art Nature-Inspired Optimization Algorithms (NIOAs) clustering approaches is conducted. According to the experimental study, the suggested method presents competitive results in terms of precision, uniformity, and robustness of the segmented outcomes contrasted to state-of-art NIOA-based clustering approaches.

show abstract

Fast K-Means Color Image Clustering With Normalized Distance Values

Cited by 3 publications

References 13 publications

Evaluating the attributes of remote sensing image pixels for fast k-meansclustering

Evaluating the attributes of remote sensing image pixels for fast k-meansclustering

Roof detection on satellite images

A novel Histogram Image Clustering Approach using Enhanced Firefly Algorithm with K-means and expanded exploitation of Aquila Optimizer

Contact Info

Product

Resources

About