GaitMDF: Gait recognition via motion deformation field modeling and knowledge transfer

Silhouette-based gait recognition is a popular biometric modality because of its non-intrusive nature and its ability to capture identity information from a distance and at low resolution. However, cross-view gait recognition presents significant challenges because the appearance changes drastically across different viewpoints. Existing gait datasets are sparse in view coverage, limiting the generalization of models to novel views. The evaluation of current state-of-the-art gait recognition models is confined to the views present in the training dataset, posing challenges for practical application scenarios. In this paper, we propose a three-dimensional (3D) human-model-guided view-morphing framework to generate high-quality gait silhouettes for arbitrary views, enhancing gait recognition performance for novel views unseen during training. By leveraging the body topology of the 3D human model, unlike existing GAN-based or deformation-based methods that rely only on two-dimensional images, the proposed method enables fine control over the output views and achieves plausible deformations, even for self-occluded body parts, through the supervision of the pseudo ground-truth deformation field computed from the 3D model. To enhance the realism of silhouette generation, we use clothed silhouettes alongside skin silhouettes rendered from the 3D model (e.g., the SMPL) to train the deformation estimation network to account for clothing and accessories in the final output. We evaluated our approach on the OU-MVLP, CASIA-B, and CASIA-E datasets, demonstrating significant improvements in gait recognition performance, particularly in view-limited scenarios. For example, recovering 6.47% rank-1 recognition rate on OU-MVLP using three close views only and an average of 24.36% and 8.34% on CASIA-B and CASIA-E normal clothing, respectively, using a single view. Our framework increases the robustness of the cross-view problem, enhancing the applicability of gait recognition in real-world settings such as surveillance and security.

The paper presents mathematical models of functioning and numerical values of information transfer coefficients for new built-in fibre-optic TFBGs-sensors (Tilted Fiber Bragg Gratings) with tilted Bragg gratings to diagnose a complex stress-strain state inside loaded polymer composite structures. The fiber optic TFBGs sensors have the form of a continuous structured cable system, in which six unidirectional light guides with the built-in Bragg gratings are placed with a fixed mutual hexagonal arrangement in extended continuous cylindrical polymer sensor housing. Different 3D orientations of reflecting surfaces for different light guides were defined through the coordinates of non-planar normals to these surfaces. Numerical modeling of deformation fields in the elements of the fiber optic TFBGs-sensor was carried out for the calculation area composite material/built-in sensor within the linear theory of elasticity. We present color diagrams of distributions for various components of the strain field along the middle cross-section of the calculation area with corresponding simple cases of its macrostrains. Also we give numerical values of strain tensor components averaged over the area of each light guide. Further, values of strain components averaged over the light guides are used to calculate axial strains along non-planar vectors - normals to reflecting surfaces of the tilted Bragg gratings. As a result, numerical values of the desired information transfer coefficients of the fibre-optic TFBGs sensor were found taking into account given orientations of reflecting surfaces of the tilted Bragg gratings of the sensor light guides. Thus, the task of diagnosing a complex deformed state inside a loaded polymer composite structure in a local neighborhood of a built-in fiber optic TFBGs sensor is reduced to solving a system of linear algebraic equations regarding the desired six independent components of the macrodeformation tensor of this neighborhood from the measured spectra of reflections of the optical fibers of the sensor.

Considering the importance of predicting volcano eruption and the large number of volcanic peaks across the world, having enough researches in this field is vital. Based on geodetic researches, surface deformation of the earth in the form of uplift or subsidence in the volcanic region, is an indicator of the magma's movement respectively towards the crater or the exit from the tank around. In this study, using the new method of the fundamental solution, the determination of the deformation field of the Campi Flegrei volcano, taking into account the topographic effect has been considered. Method of fundamental solution is a numerical Meshless method which is used to solve boundary value problems. No need for high computational cost in this method causes it becomes an effective tool in solving many problems in a variety of fields. To verify the efficiency of the method, the resulted displacement field was compared with the Interferometric Synthetic Aperture Radar (InSAR) observations at the Italian Campi Flegrei volcano. The root mean square error was less than 2 mm which indicates a good satisfactory with the displacement field resulting from the observations.