首页> 外文会议>Annual Conference of the Society for Imaging Science and Technology >Color Correction Judgements of Digital Images by Experts and Naive Observers
【24h】

Color Correction Judgements of Digital Images by Experts and Naive Observers

机译:专家和天真观察员的数字图像颜色校正判断

获取原文

摘要

We have studied the preference judgement of pictorial images with two types of population: image experts and naive observers. We used eight images of indoor and outdoor scenes. The images were first presented to image experts working in pre-press companies. We asked them to improve the images the way they preferred and to comment on their manipulations. They saved the intermediate enhanced versions they judged important. In the second part of this study, we showed those different versions of the same image produced by experts to naive observers. We used a pair comparison protocol in which all possible pairs were presented. The question was "which image do you prefer?" Observers told us their criteria for choosing an image. Several results emerged from this study. To enhance an image, an expert divides it into large zones of interest, which mainly correspond to natural colors. Likewise, when judging an image for preference, naive observers principally focus on natural colors like sky, skin or grass when present. Both experts and naives do not focus on parts if no memory colors are associated. The segmentation process into zones permits to first adjust the illuminant and then to correct the other parts with respect to the plausibility and coherence of the whole image. Further studies are necessary to correlate the chromatic signal in the retina to those data.
机译:我们研究了用两种类型的人口描绘的偏好判断:图像专家和天真观察者。我们使用了八个室内和室外场景的图像。第一个图像首先呈现给在新闻前公司工作的图像专家。我们要求他们以他们竞选的方式改进图像,并评论其操纵。他们保存了他们判断重要的中间增强版本。在本研究的第二部分,我们向幼稚观察者展示了由专家生产的相同图像的不同版本。我们使用了一对比较协议,其中呈现了所有可能对的。问题是“您更喜欢哪种形象?”观察员告诉我们他们选择图像的标准。这项研究中出现了几个结果。为了增强图像,专家将其划分为大型感兴趣的区域,主要与自然颜色相对应。同样地,当判断偏好的图像时,天真观察者主要专注于当前天空,皮肤或草等自然色彩。如果没有相关的记忆颜色,专家和天真都不关注零件。分割过程进入区域允许首先调整光源,然后校正另一个部分相对于整个图像的符号和相干性。进一步的研究是必要的,以将视网膜中的色信号与那些数据相关联。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号