arXiv subjects

Shihao Jin

Publications and source records attributed to Shihao Jin.

5 recordsLinked to original sources

arXiv2026

HINT: Human-Intent Inception for Long-Horizon Robot Manipulation

Humans can perform complex manipulations given a simple intent through an overall instruction, while continuously adapting to evolving visual observations. However, current vision-language action (VLA) models and other action policies struggle to realize this high-level intelligent behavior under dense, evolving visual inputs and sparse language guidance. Visual correlations can then dominate semantic intent, leading actions to follow visual shortcuts rather than human goals. We present HINT (Human-INTent INcepTion), an agentic framework inspired by the human manipulation principles: semantic intent changes sparsely at manipulation-pattern transitions, whereas continuous control primarily depends on the evolving object-hand relationship. HINT invokes semantic reasoning only at pattern transitions to resolve the current subtask and target, then maintains this commitment through multi-view grounding and visual tracking. We explore two visual interfaces-image-space semantic highlighting and attention-prior injection-to communicate the tracked intent to the action policy without introducing additional trainable parameters into the foundation action model. Experiments across three long-horizon tasks and out-of-distribution variants show that HINT substantially improves intent understanding, task progress, and end-to-end success across two foundation policies while preserving low-latency control. Project page: https://robot-hint.github.io/

cs.RO↗

arXiv2026

Linear Coding of LTI Sources Over Vector Gaussian Channels: A Majorization Approach

We study the design of linear time-invariant (LTI) encoder-decoder pairs for transmitting the state of a discrete-time LTI vector source over power-constrained parallel Gaussian channels with feedback. Two types of power constraints are considered. Under individual subchannel power constraints, a necessary and sufficient condition for designing an encoder-decoder pair that achieves bounded estimation error covariance (EEC) is established via two coupled majorization inequalities involving the subchannel signal-to-noise ratios and the antistable poles of the source. Under total channel power constraint, we derive the minimum total power required for a feasible encoder-decoder design by exploiting partial-order progamming under majorization order. An analytical optimal power allocation is obtained for the case of equal noise variances, which admits a water-filling interpretation; for general noise case, a sequential water-filling algorithm is developed. Our results reveal that the difficulty of transmitting a discrete-time LTI source via LTI coding is governed not only by its topological entropy, but also by the evenness of the log-magnitudes of its antistable poles. The design methods for feasible encoder-decoder pairs are also provided.

cs.IT↗

arXiv2026

fMRI2Face: A Full-HD fMRI-Video Dataset and Geometry-Guided Neural Decoding Framework for Dynamic Human Face Reconstruction

Reconstructing dynamic human faces from brain activity provides a powerful way to study how the mind perceives identity, expression, and facial motion. However, progress in fMRI-based face decoding has been limited by scarce controlled, high-resolution neural datasets and by methods that struggle to recover both identity-specific appearance and time-varying facial dynamics. We present fMRI-Face, the first fMRI dataset paired with controllable full-HD digital human facial videos rendered at 1920$\times$1080 resolution. During scanning, participants watched photorealistic, background-free facial videos with controlled identity, expression, and head pose, while fMRI activity was recorded. The resulting dataset contains 62,856 paired fMRI-video samples, providing a structured resource for studying dynamic face perception and reconstruction. Building on this dataset, we propose fMRI2Face, a geometry-guided neural video decoding framework for reconstructing facial videos from fMRI signals. fMRI2Face derives two complementary neural controls from brain activity: Brain-derived Appearance Context, which captures global identity-related visual attributes, and Morphable 3D Facial Control, which provides explicit geometry-aware guidance for pose, expression, and non-rigid facial dynamics. These controls are integrated through Neural-Controlled Video Diffusion with auxiliary latent completion, enabling high-fidelity facial video reconstruction directly from brain activity. Experiments show that fMRI2Face consistently improves reconstruction fidelity, identity preservation, facial geometry, and motion consistency over representative neural decoding baselines. Together, fMRI-Face and fMRI2Face establish a controlled platform for studying dynamic face perception and provide a new benchmark for fMRI-based digital human reconstruction.

cs.CV↗

arXiv2022

A Virtual Point Light Generation Method in Close-Range Area

This paper proposes a new hybrid algorithm for sampling virtual point light (VPL). The indirect lighting calculation of the scene is used to distribute the VPL reasonably. In the process of generating VPL, we divide the scene into two parts according to the camera position and orientation. The close-range part: the part that the camera pays attention to. The distant-range part: the part that the camera does not pay attention to or rarely pays attention to. For the close-range part, we use a patch-based vPL sampling method to distribute the VPL as evenly as possible on the patch in the near-field area; for the distant-range part, we use sparse instant radiosity (IR) for sampling. It turns out that, in contrast to conventional multiple instant radiance Compared with the VPL generation algorithm, the method proposed in this paper can greatly improve the quality of the final result graph when the number of VPLs is the same; Under the same rendering quality, the rendering speed can be greatly improved.

cs.GR↗

arXiv2019

Joint analysis of clinical risk factors and 4D cardiac motion for survival prediction using a hybrid deep learning network

In this work, a novel approach is proposed for joint analysis of high dimensional time-resolved cardiac motion features obtained from segmented cardiac MRI and low dimensional clinical risk factors to improve survival prediction in heart failure. Different methods are evaluated to find the optimal way to insert conventional covariates into deep prediction networks. Correlation analysis between autoencoder latent codes and covariate features is used to examine how these predictors interact. We believe that similar approaches could also be used to introduce knowledge of genetic variants to such survival networks to improve outcome prediction by jointly analysing cardiac motion traits with inheritable risk factors.

q-bio.QM↗