The Architectural Design of Deblocking Filter for Image Enhancement in the Diagnosis of Cancer

Abstract

Objective: This method is to reduce the risk of imprecise diagnosis associated with poor-quality CT images, this<br />paper presents a new technique designed to enhance the quality of medical CT images. The main objective is to improve<br />the appearance of CT images in order to obtain better visual interpretation and analysis, which is expected to ease the<br />diagnosis process. The proposed technique involves applying a deblocking filter is to enhance the visual nature of a<br />picture by diminishing the blocking artifacts. The appearance of a picture isn't clear while an antique happens. The<br />proposed deblocking filter calculation gives a strategy to expel the ancient rarities by smoothing the sharp edges of a<br />picture. Methods: With a specific goal to lessen the quantity of information access, multifaceted nature and consequently<br />to upgrade the proficiency, a six-staged pipelined structure for picture pixels are proposed. Besides, to enhance the<br />subjective and target nature of a picture the deblocking filter performs identification of the antique at the coded square<br />limits and weakens them by applying a chose filter. Result: The proposed algorithm is implemented in HDL using Xilinx<br />FPGA. The input image is converted into decimal pixel values using Matlab and this value is used as the input in HDL.<br />The proposed algorithm is compared with other blocking algorithms. Conclusion: To design an effective deblocking<br />filter with low cost, low complexity and high intensity, pipeline based systems are used. In addition to that the number<br />of memory accesses and timing efficiency also be reduced using this method. The deblocking filtering operations can<br />also easily perform in parallel on multiple processors by using six-stage of pipelined, two-line deblocking filter. The<br />parameter mean, variance, standard deviation, resolution, contrast and PSNR values are compared with the previous<br />method. Hence it shows the implementation of deblocking filter using pipelining is more efficient than others.