Search

Probabilistic human motion prediction aims to forecast multiple possible future movements from past observations. While current approaches report high diversity and realism, they often generate motions with undetected limb stretching and jitter. To address this, we introduce SkeletonDiffusion, a latent diffusion model that embeds an explicit inductive bias on the human body within its architecture and training. Our model is trained with a novel nonisotropic Gaussian diffusion formulation that aligns with the natural kinematic structure of the human skeleton. Results show that our approach outperforms conventional isotropic alternatives, consistently generating realistic predictions while avoiding artifacts such as limb distortion. Additionally, we identify a limitation in commonly used diversity metrics, which may inadvertently favor models that produce inconsistent limb lengths within the same sequence. SkeletonDiffusion sets a new benchmark on three real-world datasets, outperforming various baselines across multiple evaluation metrics. Visit our project page: https://ceveloper.github.io/publications/skeletondiffusion/

Stylizing a dynamic scene based on an exemplar image is critical for various real-world applications, including gaming, filmmaking, and augmented and virtual reality. However, achieving consistent stylization across both spatial and temporal dimensions remains a significant challenge. Most existing methods are designed for static scenes and often require an optimization process for each style image, limiting their adaptability. We introduce ZDySS, a zero-shot stylization framework for dynamic scenes, allowing our model to generalize to previously unseen style images at inference. Our approach employs Gaussian splatting for scene representation, linking each Gaussian to a learned feature vector that renders a feature map for any given view and timestamp. By applying style transfer on the learned feature vectors instead of the rendered feature map, we enhance spatio-temporal consistency across frames. Our method demonstrates superior performance and coherence over state-of-the-art baselines in tests on real-world dynamic scenes, making it a robust solution for practical applications.

Neural implicit surfaces can be used to recover accurate 3D geometry from imperfect point clouds. In this work, we show that state-of-the-art techniques work by minimizing an approximation of a one-sided Chamfer distance. This shape metric is not symmetric, as it only ensures that the point cloud is near the surface but not vice versa. As a consequence, existing methods can produce inaccurate reconstructions with spurious surfaces. Although one approach against spurious surfaces has been widely used in the literature, we theoretically and experimentally show that it is equivalent to regularizing the surface area, resulting in over-smoothing. As a more appealing alternative, we propose DiffCD, a novel loss function corresponding to the symmetric Chamfer distance. In contrast to previous work, DiffCD also assures that the surface is near the point cloud, which eliminates spurious surfaces without the need for additional regularization. We experimentally show that DiffCD reliably recovers a high degree of shape detail, substantially outperforming existing work across varying surface complexity and noise levels. Project code is available at https://github.com/linusnie/diffcd.

3D scene stylization extends the work of neural style transfer to 3D. A vital challenge in this problem is to maintain the uniformity of the stylized appearance across multiple views. A vast majority of the previous works achieve this by training a 3D model for every stylized image and a set of multi-view images. In contrast, we propose a novel architecture trained on a collection of style images that, at test time, produces real time high-quality stylized novel views. We choose the underlying 3D scene representation for our model as 3D Gaussian splatting. We take the 3D Gaussians and process them using a multi-resolution hash grid and a tiny MLP to obtain stylized views. The MLP is conditioned on different style codes for generalization to different styles during test time. The explicit nature of 3D Gaussians gives us inherent advantages over NeRF-based methods, including geometric consistency and a fast training and rendering regime. This enables our method to be useful for various practical use cases, such as augmented or virtual reality. We demonstrate that our method achieves state-of-the-art performance with superior visual quality on various indoor and outdoor real-world data., Comment: GCPR 2024

It is by now a well known fact in the graph learning community that the presence of bottlenecks severely limits the ability of graph neural networks to propagate information over long distances. What so far has not been appreciated is that, counter-intuitively, also the presence of strongly connected sub-graphs may severely restrict information flow in common architectures. Motivated by this observation, we introduce the concept of multi-scale consistency. At the node level this concept refers to the retention of a connected propagation graph even if connectivity varies over a given graph. At the graph-level, multi-scale consistency refers to the fact that distinct graphs describing the same object at different resolutions should be assigned similar feature vectors. As we show, both properties are not satisfied by poular graph neural network architectures. To remedy these shortcomings, we introduce ResolvNet, a flexible graph neural network based on the mathematical concept of resolvents. We rigorously establish its multi-scale consistency theoretically and verify it in extensive experiments on real world data: Here networks based on this ResolvNet architecture prove expressive; out-performing baselines significantly on many tasks; in- and outside the multi-scale setting.

Neural implicit representations have become a popular choice for modeling surfaces due to their adaptability in resolution and support for complex topology. While previous works have achieved impressive reconstruction quality by training on ground truth point clouds or meshes, they often do not discuss the data acquisition and ignore the effect of input quality and sampling methods during reconstruction. In this paper, we introduce a method that directly digests depth images for the task of high-fidelity 3D reconstruction. To this end, a simple sampling strategy is proposed to generate highly effective training data, by incorporating differentiable geometric features computed directly based on the input depth images with only marginal computational cost. Due to its simplicity, our sampling strategy can be easily incorporated into diverse popular methods, allowing their training process to be more stable and efficient. Despite its simplicity, our method outperforms a range of both classical and learning-based baselines and demonstrates state-of-the-art results in both synthetic and real-world datasets., Comment: 14 pages, GCPR 2024 paper

We present a novel neural implicit shape method for partial point cloud completion. To that end, we combine a conditional Deep-SDF architecture with learned, adversarial shape priors. More specifically, our network converts partial inputs into a global latent code and then recovers the full geometry via an implicit, signed distance generator. Additionally, we train a PointNet++ discriminator that impels the generator to produce plausible, globally consistent reconstructions. In that way, we effectively decouple the challenges of predicting shapes that are both realistic, i.e. imitate the training set's pose distribution, and accurate in the sense that they replicate the partial input observations. In our experiments, we demonstrate state-of-the-art performance for completing partial shapes, considering both man-made objects (e.g. airplanes, chairs, ...) and deformable shape categories (human bodies). Finally, we show that our adversarial training approach leads to visually plausible reconstructions that are highly consistent in recovering missing parts of a given object.

Histogram Equalization (HE) is a popular method for contrast enhancement. Generally, mean brightness is not conserved in Histogram Equalization. Initially, Bi-Histogram Equalization (BBHE) was proposed to enhance contrast while maintaining a the mean brightness. However, when mean brightness is primary concern, Minimum Mean Brightness Error Bi-Histogram Equalization (MMBEBHE) is the best technique. There are several implementations of Histogram Equalization on FPGA, however to our knowledge MMBEBHE has not been implemented on FPGAs before. Therefore, we present an implementation of MMBEBHE on FPGA.

We report the electrochemical performance of nanostructures of Na$_{0.66}$V$_4$O$_{10}$ as cathode material for rechargeable batteries. The Rietveld refinement of room temperature x-ray diffraction pattern shows the monoclinic phase with C2/m space group. The cyclic voltammetry curves of prepared half-cells exhibit redox peaks at ~3.1 and 2.6~V, which are due to two-phase transition reaction between V$^{5+/4+}$ and can be assigned to the single step deintercalation/intercalation of Na-ion. We observe a good cycling stability with specific discharge capacity (measured vs. Na$^+$/Na) between 80 ($\pm$2) and 30 ($\pm$2) mAh g$^{-1}$ at a current density 3 and 50~mA g$^{-1}$, respectively. The electrochemical performance of Na$_{0.66}$V$_4$O$_{10}$ electrode was also tested with Li anode, which showed higher capacity, but decay faster than Na. Using density functional theory, we calculate the Na vacancy formation energies; 3.37~eV in the bulk of the material and 2.52~eV on the (100) surface, which underlines the importance of nanostructures., Comment: submitted

The structure of the network has great impact on its traffic dynamics. Most of the real world networks follow the heterogeneous structure and exhibit scale-free feature. In scale-free network, a new node prefers to connect with hub nodes and the network capacity is curtailed by smaller degree nodes. Therefore, we propose rewiring a fraction of links in the network, to improve the network transport efficiency. In this paper, we discuss some efficient link rewiring strategies and perform simulations on scale-free networks, confirming the effectiveness of these strategies. The rewiring strategies actually reduce the centrality of the nodes having higher betweenness centrality. After the link rewiring process, the degree distribution of the network remains the same. This work will be beneficial for the enhancement of network performance., Comment: 14 pages

We report the synthesis, physical properties and electrochemical performance of Zn substituted Na$_{0.44}$Mn$_{1-x}$Zn$_x$O$_2$ ($x=$ 0 -- 0.02) nanostructures as cathode in Na-ion batteries for energy storage applications. These samples stabilize in the orthorhombic structure and the morphology is found to be slab like with 100 -- 200~nm width and few micrometer of length. The resistivity measurements show highly insulating nature for all the samples, where the activation energy decreases with increasing Zn concentration indicating more defect levels in the band gap. The cyclic voltammogram (CV) shows reversible oxidation and reduction peaks, which clearly shift towards higher/lower potentials with increasing Zn concentration up to 2\%. We observed the specific capacity of about 100~mAh/g at current density of 4 mA/g and improved cycle life for Zn substituted $x=$ 0.005 sample. However, with further increasing the Zn concentration ($x=$ 0.02), the specific capacity decreases, which can be a manifestation of the decreased number of reduction peaks in the CV data., Comment: accepted for publication in Ceramics International