The tumor microenvironment (TME) is now recognized as an important participant of tumor progression. As the most abundant extracellular matrix component in TME, collagen plays an important role in tumor development. The imaging study of collagen morphological feature in TME is of great significance for understanding the state of tumor. Multiphoton microscopy (MPM), based on second harmonic generation (SHG) and two-photon excitation fluorescence (TPEF), can be used to monitor the morphological changes of biological tissues without labeling. In this study, we used MPM for large-scale imaging of early invasive breast cancer from the tumor center to normal tissues far from the tumor. We found that there were significant differences in collagen morphology between breast cancer tumor boundary, near tumor transition region and normal tissues far from the tumor. Furthermore, the morphological feature of eight collagen fibers was extracted to quantify the variation trend of collagen in three regions. These results may provide a new perspective for the optimal negative margin width of breast-conserving surgery and the understanding of tumor metastasis.

Accurate thickness recognition plays a vital role in safeguarding the structural reliability of critical assets. Pulse eddy current testing (PECT), as a non-destructive method that is both non-contact and insensitive to surface coatings, provides an efficient pathway for this purpose. Nevertheless, the complex, nonstationary, and nonlinear characteristics of PECT signals make it difficult for conventional models to jointly capture localized high-frequency patterns and long-range temporal dependencies, thereby constraining their prediction performance. To overcome these issues, we introduce a novel deep learning framework, multi-scale residual dilated convolution, and bidirectional long short-term memory with a squeeze-and-excitation mechanism (MRDC-BiLSE) for PECT time series analysis. The architecture integrates a multi-scale residual dilated convolution block. By combining dilated convolutions with residual connections at different scales, this block captures structural patterns across multiple temporal resolutions, leading to more comprehensive and discriminative feature extraction. Furthermore, to better exploit temporal dependencies, the BiLSTM-SE module combines bidirectional modeling with a squeeze-and-excitation mechanism, resulting in more discriminative feature representations. Experiments on experimental PECT datasets confirm that MRDC-BiLSE surpasses existing methods, showing applicability for real-world thickness recognition.