Fusion of multimodal medical images using Daubechies complex wavelet transform – A multiresolution approach
Introduction
Biomedical image processing is a rapidly growing area of research from last two decades. Availability of numerous kinds of biomedical sensors has increased the interest of researchers and scientists in this field. X-ray, ultrasound, magnetic resonance imaging (MRI) and computed tomography (CT) are a few examples of biomedical sensors. These sensors are used for extracting clinical information, which are generally complementary in nature. For example, X-ray is widely used in detecting fractures and abnormalities in bone position, CT is used in tumor and anatomical detection and MRI is used to obtain information among tissues. Similarly, other functional imaging techniques like functional magnetic resonance imaging (fMRI), positron emission tomography (PET), single positron emission computed tomography (SPECT) provide functional and metabolic information. Hence, one can easily conclude that none of these modalities is able to carry all relevant information in a single image. Therefore, multimodal fusion is required to obtain all possible relevant information in a single composite image. Medical image fusion [1] is the process of combining and merging complementary information into a single image from two or more source images which facilitate in more precise diagnosis and better treatment. Fused image provides higher accuracy and reliability by removing redundant information. Some medical applications of image fusion are found in radiology, molecular and brain imaging, oncology, diagnosis of cardiac diseases, neuroradiology and ultrasound [2], [3], [4], [5], [6], [7], [8], [9], [10], etc.
There are two basic requirements for image fusion [11], [12]. First, fused image should possess all possible relevant information contained in the source images; second, fusion process should not introduce any artifact, noise or unexpected feature in the fused image.
Generally, pyramid and wavelet transforms are used for multiresolution image fusion. A detailed literature review on image fusion can be found in Section 2 (Background and Literature) of this paper. Real valued wavelet transform based fusion methods suffer from shift sensitivity [13] and lack of phase information [14]. Therefore, we have used Daubechies complex wavelet transform (DCxWT) [15] for image fusion, which is approximately shift invariant and provides phase information through its imaginary coefficients.
In the present work, we have proposed a new multimodal medical image fusion method using DCxWT which is based on multiresolution principle and performed multilevel fusion over three sets of multimodal medical images using maximum selection scheme. The proposed fusion method is compared with other wavelet domain (Dual tree complex wavelet transform (DTCWT), Lifting wavelet transform (LWT), Stationary wavelet transform (SWT) and Multiwavelet transform (MWT)) and spatial domain (PCA, linear and sharp) image fusion methods. The proposed method is further compared with advanced multiresolution transform (Contourlet transform (CT) and Nonsubsampled contourlet transform (NSCT)) based image fusion [16] methods. The superiority of the proposed fusion method is validated using well known fusion metrics (entropy, edge strength , standard deviation, fusion factor and fusion symmetry) for normal and noisy cases (Gaussian, salt & pepper and speckle) [17].
Rest of the paper is organized as follows: Image fusion literature is discussed in Section 2. Constructions and properties of DCxWT are described in Section 3. Section 4 explains the proposed fusion method. Experimental results and performance evaluations are given in Sections 5 Experimental results, 6 Performance evaluation respectively. Finally, conclusions of the work are given in Section 7.
Section snippets
Background and literature
Image fusion [18] is the process of integrating all relevant and complementary information from different source images into a single composite image without introducing any artifact or noise. Image fusion can be performed at three levels – pixel level [19], feature level [20] and decision level [21]. Pixel level fusion deals with information associated with each pixel and fused image can be obtained from the corresponding pixel values of source images. In feature level fusion, source images
Daubechies complex wavelet transform
The basic equation of multiresolution theory is the scaling equationwhere ak are the coefficients. The ak can be real as well as complex valued and ∑ak = 1.
Daubechies’s wavelet bases {ψj,k(t)} in one dimension are defined through the above scaling function and multiresolution analysis of L2(R) [45]. To provide general solution, Daubechies considered ak being real valued only. The construction of complex Daubechies wavelet transform is as in [42].
The generating wavelet ψ(t) is
The proposed fusion method
The proposed method uses DCxWT for multilevel medical image fusion. A general image fusion scheme using DCxWT is shown in Fig. 2.
The first step of image fusion is to decompose source images using DCxWT. For wavelet decomposition of the source images, a proper selection of mother wavelet is vital. There is no fixed criterion for selection of mother wavelet [47]. But vanishing moment and regularity (smoothness) of wavelet should be considered to decide mother wavelet [36], [47]. Mother wavelet
Experimental results
This section gives visual representation of fusion results for the proposed method. It has been assumed that all source medical images are properly registered. Experiments have been performed over two different modality of images; CT and MR. Both of these medical imaging modalities are complementary in nature. CT images are sensitive to bone and hard tissues while MR images are more informative about soft tissues. Fusion of these two will provide a single image which will be more useful for
Performance evaluation
It is a well known fact that none of the image quality metrics can directly imply the visual quality of images, hence we have to consider both the visual representation and quantitative assessment of fused images. For evaluation and comparison of the proposed fusion method, we have considered five different fusion performance metrics [22], [51], [52], [53]: entropy, edge strength , standard deviation, fusion factor and fusion symmetry. These performance metrics are defined as below-
Conclusions
Multimodal medical image fusion plays an important role in medical diagnostics. But the real challenge is to obtain visually enhanced image through fusion process. In this paper, we have proposed a new multilevel Daubechies complex wavelet transform (DCxWT) based multimodal medical image fusion method which follows multiresolution principle. Shift invariance, multiscale edge information and availability of phase information properties of DCxWT make it suitable for image fusion. Results were
Acknowledgements
This work was supported in part by the Department of Science and Technology, New Delhi, India, under Grant No. SR/FTP/ETA-023/2009 and the University Grants Commission, New Delhi, India, under Grant No. 36-246/2008(SR). The authors are thankful to Dr. J. Tian for his fruitful discussions during revision of the manuscript.
References (53)
Information fusion in the realm of medical applications – a bibliographic glimpse at its growing appeal
Information Fusion
(2012)- et al.
Cardiac health diagnosis using data fusion of cardiovascular and haemodynamic signals
Computer Methods and Programs in Biomedicine
(2006) - et al.
Image fusion of ultrasound and MRI as an aid for assessing anatomical shifts and improving overview and interpretation in ultrasound-guided neurosurgery
International Congress Series
(2001) - et al.
Image fusion in neuroradiology: three clinical examples including MRI of Parkinson disease
Computerized Medical Imaging and Graphics
(2007) - et al.
Despeckling of medical ultrasound images using complex wavelet transform
Signal Processing
(2010) - et al.
Complex Daubechies wavelets: properties and statistical image modeling
Signal Processing
(2004) - et al.
Performance comparison of different multi-resolution transforms for image fusion
Information Fusion
(2011) - et al.
Tunable halfband-pair wavelet filter banks and application to multifocus image fusion
Pattern Recognition
(2012) Additive integration of SAR features into multispectral SPOT images by means of the a trous wavelet decomposition
ISPRS Journal of Photogrammetry & Remote Sensing
(2006)- et al.
MRI and PET image fusion by combining IHS and retina-inspired models
Information Fusion
(2010)
Image fusion by a ratio of low-pass pyramid
Pattern Recognition Letters
Multisensor image fusion using the wavelet transform
Graphical Models and Image Processing
Pixel and region based fusion with complex wavelets
Information Fusion
Complex daubechies wavelets
Journal of Applied and Computational Harmonic Analysis
Multisensor and multiresolution image fusion using the linear mixing model
Multi-focus image fusion using a bilateral gradient-based sharpness criterion
Optics Communications
A novel approach to quantitative evaluation of hyperspectral image fusion techniques
Information Fusion
Head and neck cancer: clinical usefulness and accuracy of PET/CT image fusion
Radiology
Clinical value of image fusion from MR and PET in patients with head and neck Cancer
Molecular Imaging and Biology
Image fusion using CT, MRI and PET for treatment planning, navigation and follow up in percutaneous RFA
Experimental Oncology
Multisensor fusion for atrial and ventricular activity detection in coronary care monitoring
IEEE Transactions on Biomedical Engineering
Wavelets for image fusion
Daubechies complex wavelet transform based multilevel shrinkage for deblurring of medical images in presence of noise
International Journal on Wavelets, Multiresolution and Information Processing
Cited by (287)
LRFNet: A real-time medical image fusion method guided by detail information
2024, Computers in Biology and MedicineA novel approach for seismic signal denoising using optimized discrete wavelet transform via honey badger optimization algorithm
2023, Journal of Applied GeophysicsMultimodal medical image fusion towards future research: A review
2023, Journal of King Saud University - Computer and Information SciencesA review of image fusion: Methods, applications and performance metrics
2023, Digital Signal Processing: A Review JournalCDRNet: Cascaded dense residual network for grayscale and pseudocolor medical image fusion
2023, Computer Methods and Programs in BiomedicineMDRANet: A multiscale dense residual attention network for magnetic resonance and nuclear medicine image fusion
2023, Biomedical Signal Processing and Control