Survey of contemporary trends in color image segmentation

Vantaram, Sreenath Rao; Saber, Eli

doi:10.1117/1.jei.21.4.040901

Cited by 105 publications

(65 citation statements)

References 224 publications

(263 reference statements)

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Widely used alternatives include the MeanShift algorithm [52,53] or approaches based on watershed-segmentation [54]. A recent overview of trends in image segmentation is provided in Vantaram and Saber [55].…”

Section: Introductionmentioning

confidence: 99%

Self-Guided Segmentation and Classification of Multi-Temporal Landsat 8 Images for Crop Type Mapping in Southeastern Brazil

et al. 2015

View full text Add to dashboard Cite

Only well-chosen segmentation parameters ensure optimum results of object-based image analysis (OBIA). Manually defining suitable parameter sets can be a time-consuming approach, not necessarily leading to optimum results; the subjectivity of the manual approach is also obvious. For this reason, in supervised segmentation as proposed by Stefanski et al. (2013) one integrates the segmentation and classification tasks. The segmentation is optimized directly with respect to the subsequent classification. In this contribution, we build on this work and developed a fully autonomous workflow for supervised object-based classification, combining image segmentation and random forest (RF) classification. Starting from a fixed set of randomly selected and manually interpreted training samples, suitable segmentation parameters are automatically identified. A sub-tropical study site located in São Paulo State (Brazil) was used to evaluate the OPEN ACCESSRemote Sens. 2015, 7 14483proposed approach. Two multi-temporal Landsat 8 image mosaics were used as input (from August 2013 and January 2014) together with training samples from field visits and VHR (RapidEye) photo-interpretation. Using four test sites of 15 × 15 km 2 with manually interpreted crops as independent validation samples, we demonstrate that the approach leads to robust classification results. On these samples (pixel wise, n ≈ 1 million) an overall accuracy (OA) of 80% could be reached while classifying five classes: sugarcane, soybean, cassava, peanut and others. We found that the overall accuracy obtained from the four test sites was only marginally lower compared to the out-of-bag OA obtained from the training samples. Amongst the five classes, sugarcane and soybean were classified best, while cassava and peanut were often misclassified due to similarity in the spatio-temporal feature space and high within-class variabilities. Interestingly, misclassified pixels were in most cases correctly identified through the RF classification margin, which is produced as a by-product to the classification map.

show abstract

Section: Introductionmentioning

confidence: 99%

Self-Guided Segmentation and Classification of Multi-Temporal Landsat 8 Images for Crop Type Mapping in Southeastern Brazil

et al. 2015

View full text Add to dashboard Cite

show abstract

“…Contour-based approaches usually starts with edge detection, followed by a linking process to exploit curvilinear continuity, while region-based approaches try to find partitions of the image pixels into different sets corresponding to coherent image properties such as brightness, color and texture [60]. A comprehensive survey for the contemporary trends in color segmentation technique can be found in [58].…”

Section: Color Segmentationmentioning

confidence: 99%

Customized 3D Clothes Modeling for Virtual Try-on System based on Multiple Kinects

Huang

Lee

2016

Proceedings of the 11th Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications

View full text Add to dashboard Cite

Most existing 3D virtual try-on systems put clothes designed in one environment on a human captured in another environment, which cause the mismatching color and brightness problem. And also typical 3D clothes modeling starts with manually designed 2D patterns, deforms them to fit on a human, and applies stitching to sew those patterns together. Such work usually relies on labor work.In this thesis, we describe an approach to reconstruct clothes and human that both are from the same space. With multiple Kinects, it models the 3D clothes directly out of a dressed mannequin without the need of the predefined 2D clothes patterns, and fits them to a human user. Our approach has several advantages: (i) a simple hardware setting consisted of multiple Kinects to capture a human model; (ii) 3D clothes modeling directly out of captured human figure. To the best of our knowledge, our work is the first one which separates clothes out of captured human figure; (iii) resizing of clothes adapting to any sized human user; (iv) a novel idea of virtual try-on where clothes and human are captured in the same location.iii

show abstract

“…Separating the spatial-spectral properties of an image requires a proper segmentation method that partitions the pixels into clusters or groups of similar attributes such as color, texture or shape [1]. The segmented image then can be processed to enhance, analyze, detect or classify objects of interest.…”

Section: Introductionmentioning

confidence: 99%

Thoughts on Twin Earth

Jacquette

2016

Metaphysica

View full text Add to dashboard Cite

This article presents a comprehensive study of 30 color transforms to accurately segment images of halftone prints and thus calculating the parameters of a color prediction model. The transforms are evaluated combining three metrics: the model accuracy, Otsu's discriminant, and correlation coefficients of histograms. Hierarchical cluster analysis is applied to determine the thresholds to segment the image histogram into paper, ink and mixed area. Among the 30 different transforms discussed in this article, 21 channels are of 7 color space models (RGB, CMYK, CIELAB, HSV, YIQ, YCbCr, and XYZ) and the other 9 channels are specially designed. Notable increase in model accuracy validates the segmentation accuracy and the necessity of choosing the appropriate transform. A set of 180 halftone images of different print properties (such as paper, halftone, ink and printing technology) has been used for the evaluation. It is found that, the most appropriate transform depends on the type of primary ink, but the corresponding transforms in CMYK color space model have shown consistent performance. CMYK-C, XYZ-Y and LAB-B are the best transforms for Cyan, Magenta and Yellow ink color respectively. YIQ-I and HSV-S are good candidates if a single transform is to be chosen for all primary ink colors.

show abstract

Survey of contemporary trends in color image segmentation

Cited by 105 publications

References 224 publications

Self-Guided Segmentation and Classification of Multi-Temporal Landsat 8 Images for Crop Type Mapping in Southeastern Brazil

Self-Guided Segmentation and Classification of Multi-Temporal Landsat 8 Images for Crop Type Mapping in Southeastern Brazil

Customized 3D Clothes Modeling for Virtual Try-on System based on Multiple Kinects

Thoughts on Twin Earth

Contact Info

Product

Resources

About