The use of realistic visualizations has gained considerable interest due to the proliferation of virtual reality equipment. This review is concerned with the theoretical basis, technical implementation, cognitive effects, and educational implications of using realistic visualizations. Realism can be useful for learners, but in several studies, more abstract illustrations have resulted in higher performance. Furthermore, a preference for realistic visualization has been declared as being based on misconceptions regarding the cognitive system. However, we argue that this perspective is unable to fully explain the conflicting results found in the literature. To fill this theoretical gap, we devised a model to describe and compare the various levels of realism found in visualizations. We define realism as a combination of three dimensions: geometry, shading, and rendering. By varying these dimensions, it is possible to create a variety of realistic graphics. Thus, when comparing different visualizations, the realism of each of these three dimensions needs to be considered individually. Based on this technical definition, we introduce a cognitive model of learning with realistic visualizations that includes three different stages: perception, schema construction, and testing. At these three stages, variables such as the perceptual load generated by the visualization, learner characteristics influencing how well details are processed, and test types that demand concrete or flexible representations can affect whether realism fosters or hinders learning. Using the cognitive model presented in this paper, more accurate predictions and recommendations concerning the use of realism can be formulated.

由于虚拟现实设备的激增，逼真的可视化的使用引起了相当大的兴趣。本综述涉及使用现实可视化的理论基础、技术实施、认知效果和教育意义。现实主义对学习者很有用，但在一些研究中，更抽象的插图会导致更高的表现。此外，对现实可视化的偏好已被宣布为基于对认知系统的误解。然而，我们认为这种观点无法完全解释文献中发现的相互矛盾的结果。为了填补这一理论空白，我们设计了一个模型来描述和比较可视化中发现的不同层次的真实感。我们将现实主义定义为三个维度的组合：几何、着色和渲染。通过改变这些尺寸，可以创建各种逼真的图形。因此，在比较不同的可视化时，需要单独考虑这三个维度中每个维度的真实性。基于这个技术定义，我们引入了一种具有现实可视化的学习认知模型，包括三个不同的阶段：感知、模式构建和测试。在这三个阶段，诸如可视化产生的感知负荷、影响细节处理程度的学习者特征以及需要具体或灵活表示的测试类型等变量都会影响现实主义是促进还是阻碍学习。使用本文提出的认知模型，