液晶与显示
液晶與顯示
액정여현시
CHINESE JOURNAL OF LIQUID CRYSTALS AND DISPLAYS
2014年
6期
1050-1056
,共7页
图像质量评价%显示器%色域映射%LED显示
圖像質量評價%顯示器%色域映射%LED顯示
도상질량평개%현시기%색역영사%LED현시
image quality assessment%display panel%gamut mapping%LED display
针对显示器图像颜色复现问题,提出了一种基于图像质量评价和图像分割-融合策略的色域映射算法。详细分析了各类色域映射算法的特点并选取4种算法作为基本色域映射算法。提出了一种色域映射图像质量评价算法,并选用SLIC 超像素算法对源图像分割。使用基本色域映射算法对源图像进行映射处理,并计算每个分块内各映射图像与源图像的相似度。根据每个图像分块的相似度,对基本色域映射图像进行选择综合处理并融合成最终图像。最后以 LED 显示屏和 LCD 显示器为例,对本文提出的算法和其他基本色域映射算法进行主观评价对比实验。实验结果证明本文算法在图像颜色保真效果上要明显优于其他算法,但是在计算速度上仍有待优化。
針對顯示器圖像顏色複現問題,提齣瞭一種基于圖像質量評價和圖像分割-融閤策略的色域映射算法。詳細分析瞭各類色域映射算法的特點併選取4種算法作為基本色域映射算法。提齣瞭一種色域映射圖像質量評價算法,併選用SLIC 超像素算法對源圖像分割。使用基本色域映射算法對源圖像進行映射處理,併計算每箇分塊內各映射圖像與源圖像的相似度。根據每箇圖像分塊的相似度,對基本色域映射圖像進行選擇綜閤處理併融閤成最終圖像。最後以 LED 顯示屏和 LCD 顯示器為例,對本文提齣的算法和其他基本色域映射算法進行主觀評價對比實驗。實驗結果證明本文算法在圖像顏色保真效果上要明顯優于其他算法,但是在計算速度上仍有待優化。
침대현시기도상안색복현문제,제출료일충기우도상질량평개화도상분할-융합책략적색역영사산법。상세분석료각류색역영사산법적특점병선취4충산법작위기본색역영사산법。제출료일충색역영사도상질량평개산법,병선용SLIC 초상소산법대원도상분할。사용기본색역영사산법대원도상진행영사처리,병계산매개분괴내각영사도상여원도상적상사도。근거매개도상분괴적상사도,대기본색역영사도상진행선택종합처리병융합성최종도상。최후이 LED 현시병화 LCD 현시기위례,대본문제출적산법화기타기본색역영사산법진행주관평개대비실험。실험결과증명본문산법재도상안색보진효과상요명현우우기타산법,단시재계산속도상잉유대우화。
To solve the image color reproduction problems for display devices,a gamut mapping algo-rithm based on image quality assessment and image segmentation-fusion strategy was proposed. Firstly,different sorts of gamut mapping algorithms were analyzed in detail and 4 algorithms were chosen as basic algorithms.And then,a gamut mapping image quality assessment method was pro-posed,and SLIC superpixel algorithm was used to carry out the source image segmentation.Gamut mapping was applied to the source image with the basic algorithms,and similarity was calculated within each image segment between mapped images and source image.According to the calculated similarities,each mapped image was chosen and comprehensively processed,and fused to the final image.Finally,with LED display panel and LCD display as an example,a subjective evaluation experiment was made to the proposed algorithm and other basic gamut mapping algorithms.Experi-ment result shows that the proposed algorithm is apparently better than other ones on keeping image color fidelity,and yet the calculation speed still needs to be improved.