Approximation Method by Block of 3D Model for Data Reduction

With the increase in remote work due to the impact of COVID-19, which started at the beginning of 2020, the use cases of the metaverse have expanded beyond games to include business applications. Smartphones equipped with ranging sensors are also becoming more popular, making the generation of 3D models by scanning real objects easier. As a result, data transfer of 3D models over information networks is becoming common. Although the speedup of information networks has made it possible to handle large 3D models, which were difficult in the past, using large 3D models of real objects in the metaverse is undesirable because it affects the performance and increases latency. Therefore, reducing the complexity and size of 3D models is necessary. One approach is to generate an approximate 3D shape using a combination of primitive blocks, such as LEGO bricks. In this paper, we propose an algorithm that reduces the amount of transferred data and decreases operational load in the metaverse by approximating 3D models of real objects with as few primitive blocks as possible. In addition, we evaluate the effectiveness of the proposed algorithm.