Search papers, labs, and topics across Lattice.
The paper introduces CMNet, a novel hybrid deep learning architecture for phase unwrapping in fringe projection profilometry (FPP) that combines CNNs and Mamba. CMNet leverages a new convolution weighted feature fusion SSM (Conv_WFF-SSM) module to integrate Mamba's selective state space mechanisms with multi-scale convolutional feature weighting, addressing both long-range interaction modeling and hierarchical feature preservation. The architecture also incorporates a parameter-free attention module (PFAM) and an iterative attentional feature fusion (IAFF) module to further enhance performance and robustness.
A CNN-Mamba hybrid, CMNet, achieves state-of-the-art phase unwrapping from single structured light images by fusing multi-scale convolutional features with Mamba's long-range dependency modeling.
Phase unwrapping is one of the key problems in fringe projection profilometry (FPP). These years, CNN-based and Transformer-based models are widely used in FPP phase unwrapping. Unfortunately, the limitation of CNNs in long-range modeling capabilities prevent them from effectively extracting fine-grained features in wrapping phase images, while Transformer-based models struggle with efficiently handling long-range dependencies due to their local focus or computational demands. Recent studies demonstrate that Mamba, a novel selective state space model (SSM), achieves efficient modeling of long-range dependencies by dynamically adapting its parameters based on input context, particularly in tasks such as long-sequence modeling. Inspired by this, we propose a hybrid deep learning model integrating Res-UNet with Mamba (CNN-Mamba network, CMNet) for phase unwrapping of single-frame fringe patterns. First, a new dual-branch skip connection module based on Mamba is proposed named convolution weighted feature fusion SSM (Conv_WFF-SSM) is proposed, integrating Mamba’s selective state space mechanisms with multi-scale convolutional feature weighting to simultaneously address long-range interaction modeling and hierarchical feature preservation during phase expansion. Second, a parameter-free attention module (PFAM) is introduced into the encoder and decoder to reduce information loss caused by downsampling and upsampling operations without increasing network parameters. Finally, the iterative attentional feature fusion (IAFF) module integrated into residual block to instead of the ordinary sum operation for the first time. Experiments demonstrate the validity and robustness of the proposed technique.
