Jiawei Xu

Jiawei Xu, Zexin Fan, Jian Yang et al.

Recently, Gaussian splatting has received more and more attention in the field of static scene rendering. Due to the low computational overhead and inherent flexibility of explicit representations, plane-based explicit methods are popular ways to predict deformations for Gaussian-based dynamic scene rendering models. However, plane-based methods rely on the inappropriate low-rank assumption and excessively decompose the space-time 4D encoding, resulting in overmuch feature overlap and unsatisfactory rendering quality. To tackle these problems, we propose Grid4D, a dynamic scene rendering model based on Gaussian splatting and employing a novel explicit encoding method for the 4D input through the hash encoding. Different from plane-based explicit representations, we decompose the 4D encoding into one spatial and three temporal 3D hash encodings without the low-rank assumption. Additionally, we design a novel attention module that generates the attention scores in a directional range to aggregate the spatial and temporal features. The directional attention enables Grid4D to more accurately fit the diverse deformations across distinct scene components based on the spatial encoded features. Moreover, to mitigate the inherent lack of smoothness in explicit representation methods, we introduce a smooth regularization term that keeps our model from the chaos of deformation prediction. Our experiments demonstrate that Grid4D significantly outperforms the state-of-the-art models in visual quality and rendering speed.

Jiawei Xu, Yonggeon Lee, Anthony Elkommos Youssef et al.

This study addresses the challenge of predicting post-stroke rigidity by emphasizing feature interactions through graph-based explainable AI. Post-stroke rigidity, characterized by increased muscle tone and stiffness, significantly affects survivors' mobility and quality of life. Despite its prevalence, early prediction remains limited, delaying intervention. We analyze 519K stroke hospitalization records from the Healthcare Cost and Utilization Project dataset, where 43% of patients exhibited rigidity. We compare traditional approaches such as Logistic Regression, XGBoost, and Transformer with graph-based models like Graphormer and Graph Attention Network. These graph models inherently capture feature interactions and incorporate intrinsic or post-hoc explainability. Our results show that graph-based methods outperform others (AUROC 0.75), identifying key predictors such as NIH Stroke Scale and APR-DRG mortality risk scores. They also uncover interactions missed by conventional models. This research provides a novel application of graph-based XAI in stroke prognosis, with potential to guide early identification and personalized rehabilitation strategies.

Jiawei Xu, Shigang Yue, Ruisheng Wang et al.

This paper proposed a novel method to replace the traditional mouse controller by using Microsoft Kinect Sensor to realize the functional implementation on human-machine interaction. With human hand gestures and movements, Kinect Sensor could accurately recognize the participants intention and transmit our order to desktop or laptop. In addition, the trend in current HCI market is giving the customer more freedom and experiencing feeling by involving human cognitive factors more deeply. Kinect sensor receives the motion cues continuously from the humans intention and feedback the reaction during the experiments. The comparison accuracy between the hand movement and mouse cursor demonstrates the efficiency for the proposed method. In addition, the experimental results on hit rate in the game of Fruit Ninja and Shape Touching proves the real-time ability of the proposed framework. The performance evaluation built up a promise foundation for the further applications in the field of human-machine interaction. The contribution of this work is the expansion on hand gesture perception and early formulation on Mac iPad.

Jiawei Xu, Ruisheng Wang, Shigang Yue et al.

The aim of this work is to build a cognitive model for the humanoid robot, especially, we are interested in the navigation and mapping on the humanoid robot. The agents used are the Alderbaran NAO robot. The framework is effectively applied to the integration of AI, computer vision, and signal processing problems. Our model can be divided into two parts, cognitive mapping and perception. Cognitive mapping is assumed as three parts, whose representations were proposed a network of ASRs, an MFIS, and a hierarchy of Place Representations. On the other hand, perception is the traditional computer vision problem, which is the image sensing, feature extraction and interested objects tracking. The points of our project can be concluded as the following. Firstly, the robotics should realize where it is. Second, we would like to test the theory that this is how humans map their environment. The humanoid robot inspires the human vision searching by integrating the visual mechanism and computer vision techniques.