Detailed Performance Assessment of Haar, Cosine, Slant and Hartley Transforms for Grayscale Image Colorization using Thepade‟s Transform Error Vector Rotation Algorithms of Vector Quantization

The detailed performance comparison of four orthogonal transforms for grayscale image colorization using Thepade’s Transform Error Vector Rotation(TTEVR) algorithm is done here with five different color pallet sizes. The Thepade’s transform error vector rotation algorithms use binary numbers represented by four error vector sequences (Haar, Cosine, Slant & Hartley). The proposed technique uses vector quantization to generate a color pallet to color the grayscale image. The proposed technique has two stages. The first stage uses source (color) image from which color traits need to be taken is used to generate color pallet using Thepade’s transform error vector rotation algorithms. In the second stage colors are transferred to a arget (grayscale) image using generated color pallet. There exist no objective criteria for qualitative analysis of performance evaluation of the colorization quality of proposed TTEVR for the orthogonal transforms alias Cosine, Haar, Slant and Hartley is done here to find better transform to be used in TTEVR based grayscale image colorization, hence the grayscale version of original color image is recolored using proposed technique and the mean squared error between original color image and recolored image is used as quality comparison criteria. Experimentation is done on 15 different images for five different color pallet sizes in RGB. The proposed techniques are compared with existing colorization technique KEVR. KEVR performs better than the proposed techniques in RGB color space. Keywords— TCEVR, THEVR, THtEVR, TSlEVR, color pallet