Determinant based Fully Automatic one Scan Adaptive Image Scaling Algorithm

Image Scaling is defined as the resizing of images either by means of up-scaling or down-scaling. Various Algorithm such as Nearest-neighbor interpolation, Bilinear and bicubic algorithms, Sinc and Lanczos resampling, Box sampling, Mipmap, Fourier-transform methods, Edge-directed interpolation, hqx, Vectorization and Deep convolutional neural networks are available to perform image scaling. Carlo Arcelli et al. proposed a fully automatic one scan adaptive image scaling algorithm in which expanded pixel is found by the sum of weighted neighboring pixel. This paper proposes a same fully automatic one scan adaptive image scaling algorithm in which expanded pixel is found by the determinant of weighted neighboring pixel instead of summation. Finally both the existing and proposed algorithm is compared by means of PSNR and SSIM