Search papers, labs, and topics across Lattice.
This paper introduces FSC&NDF, an active contour model (ACM) that integrates fuzzy superpixel centers (FSCs) and a nonlinear diffusion filter (NDF) to address limitations of traditional ACMs, such as manual contour initialization and difficulty processing complex images. The method uses YOLOv9 to locate superpixels of interest and sets their joint boundaries as the initial contour, then employs improved fuzzy superpixel clustering to extract image features and integrates these into the energy function. Experimental results demonstrate that FSC&NDF achieves higher FPS, AP, AP50, and APM compared to mainstream deep learning algorithms, indicating improved performance in instance segmentation.
By combining YOLOv9-guided superpixels, fuzzy clustering, and nonlinear diffusion, FSC&NDF achieves instance segmentation with higher FPS and AP than mainstream deep learning, challenging the dominance of deep learning in this domain.
The significant weaknesses of the active contour model (ACM) are the manual setting of contour and the inability to process images with complex information, which limits its efficiency and application scope. In this article, an ACM, called FSC&NDF, is combined with fuzzy superpixel centers (FSCs) and nonlinear diffusion filter (NDF) to solve the above two problems simultaneously. YOLOv9 is adopted to locate the superpixels of interest; the joint boundaries of these superpixels are set as the initial contour, which is close to the morphological features of the target. Improved fuzzy superpixel clustering is applied to extract image features and yield superpixel centers, and the clusters are integrated into the main body of the energy function, NDF module further enhances boundary positioning and suppresses noise. In addition, the proposed connection mechanism makes it possible to convert object detection to instance segmentation. Experimental results show that FSC&NDF overcomes the limitations of previous ACMs in all aspects and its FPS, AP, AP50, and $\mathrm {AP}_{M}$ are higher than mainstream deep learning algorithms. The platform experiment based on the telecentric lens further proves the practicality of FSC&NDF.
