Dual-Space Augmented Intrinsic-LoRA for Wind Turbine Segmentation
This addresses the domain-specific challenge of accurate wind turbine blade segmentation for automated damage detection systems, representing an incremental improvement.
The paper tackles the problem of wind turbine blade image segmentation by extending Intrinsic LoRA with a dual-space augmentation strategy combining image-level and latent-space augmentations, achieving state-of-the-art segmentation accuracy.
Accurate segmentation of wind turbine blade (WTB) images is critical for effective assessments, as it directly influences the performance of automated damage detection systems. Despite advancements in large universal vision models, these models often underperform in domain-specific tasks like WTB segmentation. To address this, we extend Intrinsic LoRA for image segmentation, and propose a novel dual-space augmentation strategy that integrates both image-level and latent-space augmentations. The image-space augmentation is achieved through linear interpolation between image pairs, while the latent-space augmentation is accomplished by introducing a noise-based latent probabilistic model. Our approach significantly boosts segmentation accuracy, surpassing current state-of-the-art methods in WTB image segmentation.