Jiahua Li

Jian-Qing Zheng, Yuanhan Mo, Yang Sun et al.

In medical imaging, the diffusion models have shown great potential for synthetic image generation tasks. However, these approaches often lack the interpretable connections between the generated and real images and can create anatomically implausible structures or illusions. To address these limitations, we propose the Deformation-Recovery Diffusion Model (DRDM), a novel diffusion-based generative model that emphasises morphological transformation through deformation fields rather than direct image synthesis. DRDM introduces a topology-preserving deformation field generation strategy, which randomly samples and integrates multi-scale Deformation Velocity Fields (DVFs). DRDM is trained to learn to recover unrealistic deformation components, thus restoring randomly deformed images to a realistic distribution. This formulation enables the generation of diverse yet anatomically plausible deformations that preserve structural integrity, thereby improving data augmentation and synthesis for downstream tasks such as few-shot learning and image registration. Experiments on cardiac Magnetic Resonance Imaging and pulmonary Computed Tomography show that DRDM is capable of creating diverse, large-scale deformations, while maintaining anatomical plausibility of deformation fields. Additional evaluations on 2D image segmentation and 3D image registration tasks indicate notable performance gains, underscoring DRDM's potential to enhance both image manipulation and generative modelling in medical imaging applications. Project page: https://jianqingzheng.github.io/def_diff_rec/

Zibo Su, Kun Wei, Jiahua Li et al.

Speech-driven talking face synthesis (TFS) focuses on generating lifelike facial animations from audio input. Current TFS models perform well in English but unsatisfactorily in non-English languages, producing wrong mouth shapes and rigid facial expressions. The terrible performance is caused by the English-dominated training datasets and the lack of cross-language generalization abilities. Thus, we propose Multilingual Experts (MuEx), a novel framework featuring a Phoneme-Guided Mixture-of-Experts (PG-MoE) architecture that employs phonemes and visemes as universal intermediaries to bridge audio and video modalities, achieving lifelike multilingual TFS. To alleviate the influence of linguistic differences and dataset bias, we extract audio and video features as phonemes and visemes respectively, which are the basic units of speech sounds and mouth movements. To address audiovisual synchronization issues, we introduce the Phoneme-Viseme Alignment Mechanism (PV-Align), which establishes robust cross-modal correspondences between phonemes and visemes. In addition, we build a Multilingual Talking Face Benchmark (MTFB) comprising 12 diverse languages with 95.04 hours of high-quality videos for training and evaluating multilingual TFS performance. Extensive experiments demonstrate that MuEx achieves superior performance across all languages in MTFB and exhibits effective zero-shot generalization to unseen languages without additional training.

Jiahua Li, Kun Wei, Zhe Xu et al.

Long videos, characterized by temporal complexity and sparse task-relevant information, pose significant reasoning challenges for AI systems. Although various Large Language Model (LLM)-based approaches have advanced long video understanding, they still struggle to achieve both completeness and efficiency in capturing task-critical information. Inspired by human progressive visual cognition, we propose CogniGPT, a framework that leverages an interactive loop between Multi-Granular Perception Agent (MGPA) and Verification-Enhanced Reflection Agent (VERA) for efficient and reliable long video understanding. Specifically, MGPA mimics human visual divergent and focused attention to capture task-related information, while VERA verifies perceived key clues to mitigate hallucination and optimize subsequent perception strategies. Through this interactive process, CogniGPT explores a minimal set of informative and reliable task-related clues. Extensive experiments on EgoSchema, Video-MME, NExT-QA, and MovieChat datasets demonstrate CogniGPT's superiority in both accuracy and efficiency. Notably, on EgoSchema, it surpasses existing training-free methods using only 11.2 frames and achieves performance comparable to Gemini 1.5-Pro.

Jiahua Li, James T. Grist, Fergus V. Gleeson et al.

Long COVID is characterized by persistent symptoms, particularly pulmonary impairment, which necessitates advanced imaging for accurate diagnosis. Hyperpolarised Xenon-129 MRI (XeMRI) offers a promising avenue by visualising lung ventilation, perfusion, as well as gas transfer. Integrating functional data from XeMRI with structural data from Computed Tomography (CT) is crucial for comprehensive analysis and effective treatment strategies in long COVID, requiring precise data alignment from those complementary imaging modalities. To this end, CT-MRI registration is an essential intermediate step, given the significant challenges posed by the direct alignment of CT and Xe-MRI. Therefore, we proposed an end-to-end multimodal deformable image registration method that achieves superior performance for aligning long-COVID lung CT and proton density MRI (pMRI) data. Moreover, our method incorporates a novel Multi-perspective Loss (MPL) function, enhancing state-of-the-art deep learning methods for monomodal registration by making them adaptable for multimodal tasks. The registration results achieve a Dice coefficient score of 0.913, indicating a substantial improvement over the state-of-the-art multimodal image registration techniques. Since the XeMRI and pMRI images are acquired in the same sessions and can be roughly aligned, our results facilitate subsequent registration between XeMRI and CT, thereby potentially enhancing clinical decision-making for long COVID management.