Noise-Injected Nova Fractals: A Procedural Approach to Real-Time Natural and Organic Texture Generation

Procedural texture generation plays a pivotal role in computer graphics, game development, and digital art. Traditional procedural methods often face a persistent dilemma between noise and structure: gradient-based noise excels at simulating disordered natural phenomena such as clouds and terrain but lacks complex geometric structures, whereas fractals based on complex dynamical systems possess infinite geometric detail but appear overly rigid and mathematical due to their strict self-similarity. In this paper, we propose a hybrid algorithmic framework—Noise-Injected Nova Fractal (NINF). By introducing a controlled random perturbation field at each step of the generalized Newton-Raphson iteration, this method successfully breaks the rigid symmetry of the Nova fractal, generating organic textures that exhibit both macroscopic structural coherence and microscopic fluidity. Experiments demonstrate that our approach can synthesize a wide variety of natural morphologies, including biological cells, pseudo-eroded terrains, and glaciers. Despite the introduction of non-linear perturbations, the algorithm maintains a real-time frame rate of 144 FPS when generating 1024×1024 resolution height maps on an Intel UHD 630. This provides a novel method for procedural texture synthesis and real-time terrain rendering.