Image Resolution Enhancement by Discrete and Stationary ...Abstract This paper proposed a new image...

a

Technovision-2014: 1st International Conference at SITS, Narhe, Pune on April 5-6, 2014

All copyrights Reserved by Technovision-2014, Department of Electronics and Telecommunication Engineering,Sinhgad Institute of Technology and Science, Narhe, PunePublished by IJECCE (www.ijecce.org) 436

International Journal of Electronics Communication and Computer EngineeringVolume 5, Issue (4) July, Technovision-2014, ISSN 2249–071X

Image Resolution Enhancement by Discrete andStationary Wavelet Decomposition

Pravin Kumar P. RathodDepartment of Electronics and Telecommunication

Sinhgad Institute of Technology and Science, Pune, Indiapravinprathod08@gmail.com

Mrs. A. V. KulkarniAsst. Professor, Deptt. of Electronics and Telecommunication

Sinhgad Institute of Technology and Science, Pune, IndiaEmail: archanabadavekulkarni@gmail.com

Abstract – This paper proposed a new image resolutionenhancement technique by using discrete wavelet transform(DWT) and a stationary wavelet transform (SWT)decomposition. So to increase the resolution of image, the lowresolution input image is given to DWT and SWT oftechnique. The DWT and SWT are divide the input imageinto four different sub bands i.e. Low-Low (LL), Low-High(LH), High-Low (HL), High-high (HH). These sub bands arethen interpolated by using bicubic interpolation technique.The interpoalated sub bands and interpolated input imageare combine by using Inverse DWT (IDWT) to generate thehigh resolution image.

Keywords – DWT and SWT Component, Interpolation.

I. INTRODUCTION

Now day’s images are available with the differentformat. These formats have the different resolution. Theresolution of any image is the details information of thatimage. So to increase the resolutions of there are some fewtechniques are available. But to increase the resolution ofimage the best technique in bicubic interpolation. With thehelp of DWT and SWT decomposition [1]. Theinterpolation technique has been used in many imageprocessing applications are super resolution [2-3].Multiple description coding [4] and facial reconstruction[5]. To increase the resolution of image in wavelet domainis relatively a new research topic. The DWT and SWTtransforms are used to decompose. An input image intofour different sub bands like Low-Low (LL), Low-High(LH), High-Low (HL) and High-High (HH). These highfrequency sub bands are interpolated using bicubicinterpolation technique and interpolated signals arecombining by using the inverse DWT and SWT to achievethe high resolution image.

II. PROPOSE IMAGE RESOLUTIONENHANCEMENT TECHNIQUE

The main laws of image by using interpolationtechnique on high frequency component is edges, with isdue to smoothing. To increase the resolution of image,reserving the edges is important. In this work DWT andSWT are used to preserve the high frequency componentof image i.e. Low-High (LH.), High-Low (HL.), High-High (HH.) frequency component.

This proposed technique uses the one level DWT todecompose an input image into four different sub bandslike LL, LH, HL, and HH. The SWT techniquedecomposes the input image using same db1 wavelet intofour different sub bands i.e. LL, LH, HL, HH. The highfrequency output i.e. LH, Hl and HH of DWT and SWTtechnique are interpolated by using the bicubicinterpolation technique with the facto of α/2.

The output of this interpolation (The high frequencysub-bands of DWT and SWT with the interpolation factorof alpha/2) and the low resolution input image with theinterpolation factor of α/2 is combining by using inverseDWT (IDWT) to achieve the high resolution image.

The figure 1 shows the block diagram of propose imageresolution enhancement technique.

Fig.1. Block diagram of proposed system

Fig.2. Structure of wavelet decomposition

The decomposition of image after applying one levelDWT and SWT are shown in figure 2 Low-Low (L.L)Low-High (LH.), High-Low(HL.), High-High (HH.)Frequency component.

Fig.3. Structure of DWT decomposition

a

Technovision-2014: 1st International Conference at SITS, Narhe, Pune on April 5-6, 2014

All copyrights Reserved by Technovision-2014, Department of Electronics and Telecommunication Engineering,Sinhgad Institute of Technology and Science, Narhe, PunePublished by IJECCE (www.ijecce.org) 437

International Journal of Electronics Communication and Computer EngineeringVolume 5, Issue (4) July, Technovision-2014, ISSN 2249–071X

Fig.4. Structure of SWT decomposition

The figure 3 and 4 shows the structure of DWT andSWT details coefficient and approximation coefficient ofimage Low-Low(L.L) Low-High(L.H.), High-Low(H.L.),High-High(H.H.) frequency componentAlgorithm:Step 1: Take the low resolution input image.Step 2: Apply DWT on Input image.Step 3: Apply SWT on input image.Step 4: Take the Interpolation by factor α/2 of DWT &SWT output with the input.Step 5: Take the IDWT of interpolated signal.Step 6: Final Output stage.PSNR

Peak Signal to Noise Ratio (PSNR) is generally used toanalyze quality of image, sound and video files in dB(decibels). PSNR calculation of two images, one originaland an altered image, describes how far two images areequal.

III. RESULTS AND DISCUSSION

Fig.5. Input Image.

Fig.6. DWT Output.

Fig.7. SWT Output.

Fig.8. Output

Figure 5 shows the low resolution input image ofBaboon Figure 6 shows the output of DWT process.Figure7 shows the output of SWT process. Figure 8 Showsthe output of this propose technique Using DWT & SWTwith High PSNR ratio.

IV. CONCLUSION

This paper proposes image resolution enhancementtechnique by DWT and SWT decomposition by usingbicubic interpolation. By applying DWT and SWT toimage taking the interpolation and simultaneously.Interpolated low resolution image are combined by usinginverse DWT to achieve the high resolution image. ThisTechnique gives the high PSNR ratio.

REFERENCES

[1] Hasan Demirel and Gholamreza Anbarjafari “IMAGEResolution Enhancement by Using Discrete and StationaryWavelet Decomposition”, IEEE TRANSACTIONS ON IMAGEPROCESSING, VOL. 20, NO. 5, MAY 2011

[2] L. Yi-bo, X. Hong, and Z. Sen-yue, “The wrinkle generationmethod for facial reconstruction based on extraction of partitionwrinkle line features and fractal interpolation,” in Proc. 4th Int.Conf. Image Graph., Aug. 22–24, 2007, pp. 933–937.

[3] H. Demirel, G. Anbarjafari, and S. Izadpanahi, “Improvedmotionbased localized super resolution technique using discretewavelet transform for low resolution video enhancement,” inProc. 17th Eur. Signal Process. Conf., Glasgow, Scotland, Aug.2009, pp. 1097–1101.

[4] C. B. Atkins, C. A. Bouman, and J. P. Allebach, “Optimal imagescaling using pixel classification,” in Proc. Int. Conf. ImageProcess., Oct. 7–10, 2001, vol. 3, pp. 864–867.

[5] Y. Rener, J. Wei, and C. Ken, “Downsample-based multipledescription coding and post-processing of decoding,” in Proc.27th Chinese Control Conf., Jul. 16–18, 2008, pp. 253–256.

a

Technovision-2014: 1st International Conference at SITS, Narhe, Pune on April 5-6, 2014

All copyrights Reserved by Technovision-2014, Department of Electronics and Telecommunication Engineering,Sinhgad Institute of Technology and Science, Narhe, PunePublished by IJECCE (www.ijecce.org) 438

International Journal of Electronics Communication and Computer EngineeringVolume 5, Issue (4) July, Technovision-2014, ISSN 2249–071X

[6] S. Mallat, A Wavelet Tour of Signal Processing, 2nd ed. NewYork: Academic, 1999.

[7] Y. Piao, I. Shin, and H. W. Park, “Image resolution enhancementusing inter-subband correlation in wavelet domain,” in Proc. Int.Conf. Image Process., 2007, vol. 1, pp. I-445–448.

[8] W. K. Carey, D. B. Chuang, and S. S. Hemami, “Regularity-preserving image interpolation,” IEEE Trans. Image Process.,vol. 8, no. 9, pp. 1295–1297, Sep. 1999.