Resolution Enhancement of American sign language Image Using DT-CWT and EPS algorithm Pradeep kumar B P Research Scholar Jain University Bangalore, India pradi14cta@gmail.com Manjunatha M B principal AIT Tumkur, India manju.kari29@gmail.com Revana siddesh p.n pg scholar AIT Tumkur, India pnsiddesh@gmail.com Abstract Nowadays, Gesture controlled applications have become an option as the user interface to a digital machine. But usage of low-cost motion sensing images captured are low and it under-performs in low lighting conditions producing blurred images. The paper proposes hand image Resolution Enhancement techniques based on Multi Scale Decomposition and Edge Preservation Smoothing. The proposed DT-CWT and EPS algorithm images are decomposed into different sub bands and interpolated, after which sub bands are reconstructed to achieve the enhanced image. Samples of sign languages have been recorded from kinect camera at the distance 1500-2000 mm and having an instant check of its effect and again later after 5 minutes. The parameters obtained are compared to those with recorded under common conditions. Signs from different users taken by kinect camera and recorded 20 images for each signs, The quantitative analysis is done by comparing PSNR, MSE and SSIM values. The authors propose to investigate those and also the variability as a parameter for the sign language recognition process. The entire process has been done using MATLAB. Key words: EPS(edge preservation smoothing), Peak signal to noise Ratio, MSE, SSIM (Structural Similarity Index for Measurement) 1. Introduction Image enhancement in American sign language is the procedure of manipulate an image so that the resulting image is more fitting than the input image for a specific application. American sign language images are used in many fields of research. The excellence and magnitude of hand image is mainly determined by their resolution. There are different types of resolution when discussing the image in remote sensing: spatial, spectral, temporal and radiometric. Image RE in sign language in wavelet domain is a new research area, to preserve components of high frequency of the image. In the resolution enhancement of satellite images ,input image is decomposed by DT-CWT and it is shift invariant to obtain high-frequency components . The HF subbands and the low-resolution input image are interpolated using the Lanczos interpolator[1].In recent times many RE algorithms have been projected (DWT, SWT, DT-CWT). Complex wavelet transform is the most recent wavelets transform used in Resolution enhancement techniques. DWT decomposes an sign image into different sub band frequencies , namely LL, LH, HL and HH. [3]. one more wavelet transform has been used in several image processing applications is Stationary Wavelet Transform (SWT). Down sampling done for the each sign image in DWT sub frequency band causes information loss. So SWT is engaged to reduce this loss.many methods for filter design are describe for dual-tree complex wavelet transform. it demonstrate with comparatively small filters, an efficient invertible approximately logical wavelet transform can implemented using the dual-tree approach. SWT is also similar to DWT but not used for down sampling, so the sub frequency bands will have the same dimension as the input image. Dual Tree-Complex Wavelet Transform (DT-CWT) is shift invariant and directional selective. In a Multi scale Decomposition of image and Edge preserving Smoothing based image RE (DTCWT-EPS) technique is proposed which generate sharper high resolution hand image. By, the proposed technique and visual results will afore said state of art and conventional techniques for hand Image enhancement.[6] International Journal of Computer Science and Information Security (IJCSIS), Vol. 15, No. 4, April 2017 126 https://sites.google.com/site/ijcsis/ ISSN 1947-5500