Performance evaluation for image fusion technique in medical images using spatial and transform method

Saranya, G.; Devi, S. Nirmala

doi:10.1109/wispnet.2016.7566173

Cited by 4 publications

(1 citation statement)

References 4 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In the literature, many image fusion methods have been presented like discrete wavelet transform (DWT) [5] and Stationary wavelet transform [6]- [7]. However, SWT is the popular techniques for image fusion because it is shift invariant which does not allow the alteration in coefficient map if there is any small mismatch between the pixels of source image [8]. In Stationary wavelet transform based image fusion, the source images are decomposed into four equal sized images.…”

Section: Introductionmentioning

confidence: 99%

Change Detection from Remotely Sensed Images Based on Stationary Wavelet Transform

Sharma¹,

Gulati²

2017

IJECE

View full text Add to dashboard Cite

The major issue of concern in change detection process is the accuracy of the algorithm to recover changed and unchanged pixels. The fusion rules presented in the existing methods could not integrate the features accurately which results in more number of false alarms and speckle noise in the output image. This paper proposes an algorithm which fuses two multi-temporal images through proposed set of fusion rules in stationary wavelet transform. In the first step, the source images obtained from log ratio and mean ratio operators are decomposed into three high frequency sub-bands and one low frequency sub-band by stationary wavelet transform. Then, proposed fusion rules for low and high frequency sub-bands are applied on the coefficient maps to get the fused wavelet coefficients map. The fused image is recovered by applying the inverse stationary wavelet transform (ISWT) on the fused coefficient map. Finally, the changed and unchanged areas are classified using Fuzzy c means clustering. The performance of the algorithm is calculated in terms of percentage correct classification (PCC), overall error (OE) and Kappa coefficient (K c ). The qualitative and quantitative results prove that the proposed method offers least error, highest accuracy and Kappa value as compare to its preexistences.

show abstract

Section: Introductionmentioning

confidence: 99%