In this paper, the authors address the challenges of applying three-dimensional virtual worlds for collaborative work and learning, such as steep learning curve and the demands for computational and network resources. We developed a texture generation model utilizing stream processors that allows displaying large amount of meaningful content in virtual worlds, reducing the technical requirements and allowing convenient tools that simplify the use of the technology, and therefore, improve the negative learning curve effect. The authors present original methods of generating images and several tools implemented in vAcademia virtual world. A tool called Sticky Notes is presented in detail as an example. In addition, the authors provide the evaluation of the suggested model and the first result of the user evaluation.