Research And Implementation Of Document Image Quality Assessment Platform

With the vigorous development of Internet technology and the improvement of the hardware of mobile devices such as phones and tablets,the images collected by shooting documents have been widely used in various industries,but the quality of the images collected in natural scenes is uneven,which directly affects the use in subsequent scenes.By quantifying the quality of document image,and filtering images based on the results,various applications related to image processing can be optimized.The document image quality assessment system is to comprehensively evaluate the document image quality and calculate several image quality scores.By analyzing the characteristics of document images,this system designs and implements several image quality assessment methods,which are used to calculate the objective scores of the three evaluation indexes of blur,deformation,and contrast of document images,and calculate the overall score according to the actual processing effects of each method.In terms of human-computer interaction,around the score calculation function,the system realizes the statistics and visualization of the evaluation results,including the objective scoring and related parameter display of a single image,and the statistical chart and table display of objective scoring of image dataset.Moreover,this system is developed based on Python's Qt toolkit and can run on the operating platform deployed with a Python environment.The thesis introduces the design and implementation of the document image quality assessment system in detail.First,analyze functional requirements according to the application scenarios of document images and designs the overall architecture of the system.Then researched the technologies related to system development and determined the technical implementation of each functional module.Next,we further analyze the processing flow of each functional module to complete the work of coding.The image quality assessment module introduces the self-collected document image dataset,which is used for feature extraction and experimental test of various evaluation methods.Finally,test the system according to the functional requirements,analyze the experimental results and optimize the system.Experimental results show that the system's evaluation results of document image quality are basically consistent with subjective evaluation' s,and can run stably in practical applications.