In this study, we introduce a novel two-stage technique for decomposing and reconstructing facial features from sparse-view images, a task made challenging by the unique geometry and complex skin reflectance of each individual. To synthesize 3D facial models more realistically, we endeavor to decouple key facial attributes from the RGB color, including geometry, diffuse reflectance, and specular reflectance. Specifically, we design a Sparse-view Face Decomposition Model (SFDM): 1) In the first stage, we create a general facial template from a wide array of individual faces, encapsulating essential geometric and reflectance characteristics. 2) Guided by this template, we refine a specific facial model for each individual in the second stage, considering the interaction between geometry and reflectance, as well as the effects of subsurface scattering on the skin. With these advances, our method can reconstruct high-quality facial representations from as few as three images. The comprehensive evaluation and comparison reveal that our approach outperforms existing methods by effectively disentangling geometric and reflectance components, significantly enhancing the quality of synthesized novel views, and paving the way for applications in facial relighting and reflectance editing. Visit our project page for more details https://kingjg.github.io/SFDM.github.io/.

We introduce MIRReS, a novel two-stage inverse rendering framework that jointly reconstructs and optimizes explicit geometry, materials, and lighting from multi-view images. Unlike previous methods that rely on implicit irradiance fields or oversimplified ray tracing, our method begins with an initial stage that extracts an explicit triangular mesh. In the second stage, we refine this representation using a physically-based inverse rendering model with multi-bounce path tracing and Monte Carlo integration. This enables our method to accurately estimate indirect illumination effects, including self-shadowing and internal reflections, leading to a more precise intrinsic decomposition of shape, material, and lighting. To address the noise issue in Monte Carlo integration, we incorporate reservoir sampling, improving convergence and enabling efficient gradient-based optimization with low sample counts. Through both qualitative and quantitative assessments across various scenarios, especially those with complex shadows, we demonstrate that our method achieves state-of-the-art decomposition performance. Furthermore, our optimized explicit geometry seamlessly integrates with modern graphics engines supporting downstream applications such as scene editing, relighting, and material editing.

Realistic human skin rendering has been a long-standing challenge in computer graphics. Recently, biophysical-based skin rendering has received increasing attention, as it provides a more realistic skin-rendering and a more intuitive way to adjust the skin style. In this work, we present a novel heterogeneous biophysical-based volume rendering method for human skin that improves the realism of skin appearance while easily simulating various types of skin effects, including skin diseases, by modifying biological coefficient textures. Specifically, we introduce a two-layer skin representation by mesh deformation that explicitly models the epidermis and dermis with heterogeneous volumetric medium layers containing the corresponding spatially varying melanin and hemoglobin, respectively. Furthermore, to better facilitate skin acquisition, we introduced a learning-based framework that automatically estimates spatially varying biological coefficients from an albedo texture, enabling biophysical-based and intuitive editing, such as tanning, pathological vitiligo, and freckles. We illustrated the effects of multiple skin-editing applications and demonstrated superior quality to the commonly used random walk skin-rendering method, with more convincing skin details regarding subsurface scattering.

Recently, metasurfaces have demonstrated considerable potential for application across various fields, including wireless communication, Internet of Things, holographic imaging, and radar systems. With the rapid development of artificial intelligence and its application in other disciplines, metasurface design based on artificial intelligence techniques, particularly deep learning, has attracted widespread attention. In this paper, we propose a novel scheme that integrates deep neural networks with an enhanced simulated annealing algorithm to achieve intelligent design of broadband metasurfaces. Firstly, we construct and train a neural network capable of accurately predicting the amplitude and phase responses of an isotropic metasurface structure, achieving a Mean Absolute Error of 1.36° on the test set. Subsequently, we develop a novel simulated annealing algorithm that is integrated with the prediction neural network to form an intelligent optimization framework. This approach can effectively search for the solution space to accurately achieve arbitrary target phase response, with a very low Mean Squared Error calculated on the test set. Furthermore, we validate the effectiveness of the proposed model and algorithm through a practical application that demonstrates broadband dispersion characteristics within the frequency range of 9–11 Gigahertz (GHz). Additionally, the prediction model can employ transfer learning strategy to adapt to various operating frequency bands, significantly enhancing its generalization capabilities. The proposed intelligent design methodology introduces a novel pathway for rapidly designing metasurface in complex application scenarios, serving as a valuable reference for the engineering design of devices in electronics, optics, and wireless communication.

Crude oil and its emulsions seriously threaten marine ecological environment and human health. Thermal infrared remote sensing technology is an important means of optical remote sensing marine oil-spill monitoring. In this paper, based on the UAV thermal infrared radiometer, a 24-hour continuous oil-spill brightness temperature image acquisition experiment was carried out, and oil-spill BTD polar coordinate thermal model was constructed to analyze the thermal infrared response characteristics of crude oil and emulsified oil. The results show that the thermal infrared response characteristics of oil-in-water (OW) emulsified oil and seawater are difficult to distinguish, which of crude oil and water-in-oil (WO) emulsion are similar. During 11:00~14:00, there is strong thermal infrared separability between oil and seawater, and the BTD values of crude oil and WO emulsified oil are both greater than 15. During 10:30~15:00, there is a strong positive correlation between BTD and oil film thickness, R2is greater than 0.9, so it is determined that the optimal time window for oil-spill thermal infrared monitoring is 11:00~14:00. The thermal infrared remote sensing monitoring based on oil-spill BTD model and optimal time window is expected to provide new technical and method support for marine oil-spill emergencies.

The workload of real-time rendering is steeply increasing as the demand for high resolution, high refresh rates, and high realism rises, overwhelming most graphics cards. To mitigate this problem, one of the most popular solutions is to render images at a low resolution to reduce rendering overhead, and then manage to accurately upsample the low-resolution rendered image to the target resolution, a.k.a. super-resolution techniques. Most existing methods focus on exploiting information from low-resolution inputs, such as historical frames. The absence of high frequency details in those LR inputs makes them hard to recover fine details in their high-resolution predictions. In this paper, we propose an efficient and effective super-resolution method that predicts high-quality upsampled reconstructions utilizing low-cost high-resolution auxiliary G-Buffers as additional input. With LR images and HR G-buffers as input, the network requires to align and fuse features at multi resolution levels. We introduce an efficient and effective H-Net architecture to solve this problem and significantly reduce rendering overhead without noticeable quality deterioration. Experiments show that our method is able to produce temporally consistent reconstructions in 4 × 4 and even challenging 8 × 8 upsampling cases at 4K resolution with real-time performance, with substantially improved quality and significant performance boost compared to existing works.Project page: https://isaac-paradox.github.io/FuseSR/

The intermittency and randomness of wind power output have a negative impact on the stable operation of the power grid. Accurately modeling the uncertainty of wind power output is essential, and the primary method to achieve this is through scenario generation. Traditional scenario generation methods suffer from limitations such as low accuracy and high computational complexity. In this paper, a novel generation framework based on the denoising diffusion probabilistic model is presented and proposed for scenario generation of wind power. This method can overcome the limitations of traditional methods and learn the distribution of real data to generate reliable wind power scenarios. Compared to a homogeneous generative model, the proposed method shows improved performance in precisely capturing features of wind power scenarios.