The increased size of grayscale images or upscale plays a central role in various fields such as medicine, satellite imagery, and photography. This paper presents a technique for improving upscaling gray images using a new mixing wavelet generation by tensor product. The proposed technique employs a multi-resolution analysis provided by a new mixing wavelet transform algorithm to decompose the input image into different frequency components. After processing, the low-resolution input image is effectively transformed into a higher-resolution representation by adding a zeroes matrix. Discrete wavelets transform (Daubechies wavelet Haar) as a 2D matrix is used but is mixed using tensor product with another wavelet matrix’s size. MATLAB R2021b is used as the main program, and the performance metrics of upscale are evaluated in gray images. The results exhibit the supremacy of the proposed approach in terms of visual quality and quantitative metrics. The upscale by mix transform using wavelet transform provides efficient upscale quality metrics as compared with the technique that upscales image by mix transform using Slantlet transform, which has an upscaled blurred image. Also, wavelet and Slantlet upscaled Peak Signal-to-Noise Ratio (PSNRus)=79.4698 and 76.6695, respectively. This approach is often used in cryptography and computer vision.
