Architectural animation generation system based on AL-GAN algorithm

As the construction industry grows and technology advances, architectural animation (AA) plays an increasingly critical role in expressing building design and configuration. However, the generation of AA faces challenges in handling the details of architectural images and maintaining animation continuity. To this end, a new AA generation system based on the architecture learning generative adversarial network (AL-GAN) algorithm is proposed, aiming to solve the problems of insufficient detail handling, poor animation continuity, and inefficiency in traditional AA generation. The new system significantly improves the architectural image detail processing capability and animation continuity by combining the attention mechanism and the long short-term memory network. The experimental results showed that AL-GAN could get closer to the real image and process architectural colors and lines more accurately and vividly, demonstrating its superior image-processing capability. In addition, AL-GAN showed the best adaptability in different scenes, reaching 99.99% adaptability in the park scene. The research model also performed well in terms of color recognition accuracy with relatively small error values, indicating better model performance and stability. Therefore, the improved AL-GAN has better performance in the AA generation system. The study not only provides a more efficient and accurate solution for the field of AA generation but also experimentally verifies the adaptability and stability of AL-GAN in different scenes.