Gaussian mixture modeling of histograms for contrast enhancement

Lai, Yu-Ren; Chung, Kuo–Liang; Lin, Guei-Yin; Chen, Chyouhwa

doi:10.1016/j.eswa.2011.12.018

Cited by 25 publications

(12 citation statements)

References 23 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…It is useful to note that in this step the performance of the enhancement methods has the least importance since the enhancement methods should only cover a good range of contrast‐based quality and not necessarily good quality. The chosen contrast enhancement methods for comparison include: (i) Histogram equalisation (HE) [36] (ii) Mean brightness HE [17] (iii) Dualistic sub‐image HE [18] (iv) Quadrants dynamic HE [37] (v) Gaussian mixture model‐based contrast enhancement [31] (vi) Expectation maximisation based contrast enhancement [38] (vii) Mean‐based recursively separated weighted HE [26] Moreover, 11 IQA tools were used to measure the quality of different enhanced versions of the images in the dataset. It is important to note that, after evaluating different state‐of‐the‐art IQA tools, it was decided to consider only the ‘contrast‐based’ quality measures in the literature.…”

Section: Resultsmentioning

confidence: 99%

Quality assessment tool for performance measurement of image contrast enhancement methods

Abdoli

Nasiri

Brault

et al. 2019

IET Image Processing

View full text Add to dashboard Cite

Section: Resultsmentioning

confidence: 99%

Quality assessment tool for performance measurement of image contrast enhancement methods

Abdoli

Nasiri

Brault

et al. 2019

IET Image Processing

View full text Add to dashboard Cite

“…To demonstrate the performance of GMMCE, the proposed method was compared with other existing histogram based contrast enhancement methods: Histogram Equalization (HE) [8], Brightness preserving Bi-Histogram Equalization (BBHE) [9], Dualistic Sub-Image Histogram Equalization (DSIHE) [11], Mean-based Recursively Separated and Weighted Histogram Equalization (RSWHE-M with r=2 as its best-tuned recursion level) [29], Quadrants Dynamic Histogram Equalization (QDHE) [17]. For thorough comparison, the proposed GMMCE was also compared to a GMM-based method called Expectation Maximization Contrast Enhancement (EMCE) [20]. It can be useful to note that in EMCE all GMM parameters (i.e.…”

Section: Resultsmentioning

confidence: 99%

Gaussian mixture model‐based contrast enhancement

Abdoli

Sarikhani²,

Ghanbari

et al. 2015

IET Image Processing

View full text Add to dashboard Cite

show abstract

“…Finally, subhistogram matching was carried out based on the histogram equalization techniques to normalize the intensity of the second LiDAR data strip with reference to the intensity of the first LiDAR data strip. In the following sections, we follow the notation of GMM, as defined in [43] for presenting the mathematical model of radiometric normalization.…”

Section: A Overall Workflowmentioning

confidence: 99%

“…Similarly, the coefficient of joint variation cjv between two classes is commonly used to assess the variation between two , ω i and ω j ) by using (42), and the relative change of the coefficient of joint variation can be determined by (43) as follows:…”

Section: B Design Of Experiments and Evaluationmentioning

confidence: 99%

Radiometric Correction and Normalization of Airborne LiDAR Intensity Data for Improving Land-Cover Classification

Yan

Shaker

2014

IEEE Trans. Geosci. Remote Sensing

View full text Add to dashboard Cite

Radiometric correction of airborne LiDAR intensity data has been proposed based on the use of the radar (range) equation for removing the effects of attenuation due to systemand environmental-induced distortions. Although radiometric correction of airborne LiDAR intensity data has been recently investigated with results revealing improved accuracy of surface classification, there exist a few voids requiring further research effort. First, the variation of object surface characteristics (slope and aspect) plays a crucial role in modeling the recorded intensity data, and thus, the laser incidence angle is usually considered in the correction process. Nevertheless, the use of incidence angle would lead to the effects of overcorrection, particularly on those features located in steep slope. Second, line-stripping problems are usually appeared in the overlapping region of LiDAR data strips acquired by sensors configured with automatic gain control (AGC). Currently, the effects of AGC cannot be perfectly modeled due to the nondisclosure of information by the sensor manufacturers. In this paper, we attempt to fill these voids by: 1) proposing a correction mechanism using the surface slope as a threshold to select either using scan angle or incidence angle in the radar (range) equation; and 2) proposing a subhistogram matching technique to radiometrically normalize the overlapping intensity data. The proposed approaches were applied to three real airborne LiDAR data strips for experimental testing. The results showed that the coefficient of variation reached to the lowest value for most of the land-cover features with a slope threshold between 30 • and 40 • . The variance-to-mean ratio of five land-cover features was significantly reduced by 70%-82% after applying the proposed correction mechanism. In addition, the systematic noises appeared in the overlapping region were significantly reduced after radiometric correction and normalization, where the overall accuracies were improved by up to 16.5% in the results of intensity data classification. With the demonstrated improvement in intensity homogeneity, it is recommended that airborne LiDAR intensity data should be radiometrically preprocessed before performing any thematic applications.Index Terms-Airborne LiDAR, Gaussian mixture model (GMM), incidence angle, intensity, land-cover classification, landcover homogeneity, radiometric correction, radiometric normalization, subhistogram matching.

show abstract

Gaussian mixture modeling of histograms for contrast enhancement

Cited by 25 publications

References 23 publications

Quality assessment tool for performance measurement of image contrast enhancement methods

Quality assessment tool for performance measurement of image contrast enhancement methods

Gaussian mixture model‐based contrast enhancement

Radiometric Correction and Normalization of Airborne LiDAR Intensity Data for Improving Land-Cover Classification

Contact Info

Product

Resources

About