MF2ResU-Net: A Multi-Feature Fusion Deep Learning Architecture for Retinal Blood Vessel Segmentation

Segmentation of blood vessels becomes an essential step in computer aided diagnosis system for the diseases in several departments of ophthalmology, neurosurgery, oncology, cardiology, and laryngology. Aiming at the problem of insufficient segmentation of small blood vessels by existing methods, a novel method based on multi-module fusion residual neural network model (MF2ResU-Net) was proposed. In the proposed networks, to obtain refined features of vessels, three cascade connected U-Net networks were employed as main networks. To deal with the problem of over-fitting, residual paths were used in main networks. In the blocks of U-Net in MF2ResU-Net, in order to remove the semantic difference in low-level and high-level, shortcut connections were used to combine the encoder layers and decoder layers in the blocks. Furthermore, atrous spatial pyramid pooling was embedded between the encoder and decoder to achieve multi-scale feature of blood vessels. During the training of the networks, to deal with the imbalance between background and foreground, a novel joint loss function was proposed based on the dice and cost- sensitive, which could greatly reduce the impact of unbalance in classes of samples. In experiment section, two retinal datasets, DRIVE and CHASE DB1, were used to test our method, and experiments showed that MF2ResU-Net was superior to existing methods on the criteria of sensitivity (Sen), specificity (Spe), accuracy (Acc), and area under curve (AUC), the values of which are 0.8013 and 0.8102, 0.9842 and 0.9809, 0.9700 and 0.9776, and 0.9797 and 0.9837 respectively for DRIVE and CHASE DB1. The results of experiments demonstrated the effectiveness and robustness of the model in the segmentation of complex curvature and small blood vessels.

Curvature domains in V4 of macaque monkey

An important aspect of visual object recognition is the ability to perceive object shape. Two basic components of complex shapes are straight and curved contours. A large body of evidence suggests a modular hierarchy for shape representation progressing from simple and complex orientation in early areas V1 and V2, to increasingly complex stages of curvature representation in V4, TEO, and TE. Here, we reinforce and extend the concept of modular representation. Using intrinsic signal optical imaging in Macaque area V4, we find sub-millimeter sized modules for curvature representation that are organized from low to high curvatures as well as domains with complex curvature preference. We propose a possible ‘curvature hypercolumn’ within V4. In combination with previous studies, we suggest that the key emergent functions at each stage of cortical processing are represented in systematic, modular maps.

Emergent collective organization of bone cells in complex curvature fields

Individual cells and multicellular systems have been shown to respond to cell-scale curvatures in their environments, guiding migration, orientation, and tissue formation. However, it remains unclear how cells collectively explore and pattern complex landscapes with curvature gradients across the Euclidean and non-Euclidean spectra, partly owing to fabrication limitations and the lack of formal geometric considerations. Here, we show that micro-engineered substrates with controlled curvature variations induce the collective spatiotemporal organization of preosteoblasts. By leveraging mathematical surface design and a high-resolution free-form fabrication process, we exposed cells to a broad yet controlled, heterogeneous spectrum of curvature fields. We quantified curvature-induced spatial patterning at different time points and found that cells generally prefer regions with at least one negative principal curvature. We also show that multicellular cooperation enables cells to venture into unfavourably-curved territories, bridging large portions of the substrates, and collectively aligning their stress fibres. We demonstrate that this behaviour is partly regulated by cellular contractility and extracellular matrix development, underscoring the mechanical nature of curvature guidance. Our findings offer unifying perspectives on cell-geometry interactions that could be harnessed in the design of micro-engineered biomaterials, for example, for tissue engineering applications.

Automated local line rolling forming and simplified deformation simulation method for complex curvature plate of ships

Local line rolling forming is a common forming approach for the complex curvature plate of ships. However, the processing mode based on artificial experience is still applied at present, because it is difficult to integrally determine relational data for the forming shape, processing path, and process parameters used to drive automation equipment. Numerical simulation is currently the major approach for generating such complex relational data. Therefore, a highly precise and effective numerical computation method becomes crucial in the development of the automated local line rolling forming system for producing complex curvature plates used in ships. In this study, a three-dimensional elastoplastic finite element method was first employed to perform numerical computations for local line rolling forming, and the corresponding deformation and strain distribution features were acquired. In addition, according to the characteristics of strain distributions, a simplified deformation simulation method, based on the deformation obtained by applying strain was presented. Compared to the results of the three-dimensional elastoplastic finite element method, this simplified deformation simulation method was verified to provide high computational accuracy, and this could result in a substantial reduction in calculation time. Thus, the application of the simplified deformation simulation method was further explored in the case of multiple rolling loading paths. Moreover, it was also utilized to calculate the local line rolling forming for the typical complex curvature plate of ships. Research findings indicated that the simplified deformation simulation method was an effective tool for rapidly obtaining relationships between the forming shape, processing path, and process parameters.

Aesthetic Curves and Surfaces in Computer Aided Geometric Design

Aesthetic shapes are usually actualized as 3D objects represented by free-form surfaces. The main components used to achieve aesthetic surfaces are 2D and 3D curves, which are the elements most basic for determining the shapes and silhouettes of industrial products. Bézier, B-Spline and NURBS are types of flexible curves developed for various design intents. These curves, however produce complex curvature functions that may undermine the formulation of shape aesthetics. A viable solution to this problem is to formulate aesthetic curves and surfaces from well-defined curvatures to improve aesthetic design quality. This paper advocates formalizing aesthetic curve and surface theories to fill the gapmentioned above, which has existed since the 1970s. This paper begins by reviewing on fair curves and surfaces. It then extensively discusses on the technicalities of Log-Aesthetic (LA) curves and surfaces and touches on industrial design applications. These emerging LA curves have a high potential for being used as standards to generate, evaluate and reshape aesthetic curves and surfaces, thus revolutionizing efficiency in developing curve and shape aesthetics.