12 January 2026 in IEEE Transactions on Pattern Analysis and Machine Intelligence
Diffusion-based technologies have made significant strides, particularly in personalized and customized facial generation. However, existing methods struggle to achieve high-fidelity and detailed identity (ID) consistency. This is mainly due to two challenges: insufficient fine-grained control over specific facial areas and the absence of a comprehensive strategy for ID preservation that accounts for both intricate facial details and the overall facial structure. To address these limitations, we introduce ConsistentID, an innovative method crafted for diverse identity-preserving portrait generation under fine-grained multimodal facial prompts, utilizing only a single reference image. ConsistentID comprises two core components: a multimodal facial prompt generator and an ID-preservation network. The facial prompt generator combines localized facial features, facial feature descriptions, and overall facial descriptions to enhance the precision of facial detail reconstruction. The ID-preservation network, optimized with a facial attention localization strategy, ensures consistent identity preservation across facial regions. Together, these components leverage fine-grained multimodal identity information to improve identity preservation accuracy significantly. To drive ConsistentID’s training, we propose a fine-grained portrait dataset, FGID, with over 500,000 facial images, offering greater diversity and comprehensiveness than existing public facial datasets. Experimental results substantiate that our ConsistentID achieves exceptional precision and diversity in personalized facial generation, surpassing existing methods in the MyStyle dataset. In addition, although ConsistentID introduces more multimodal ID information, it still maintains rapid inference speed during the generation process.
21 February 2025 in IEEE Transactions on Instrumentation and Measurement
Sign assessment in Huntingtons disease (HD) holds significant importance for its treatment. However, most of the existing methods to assess motor signs have shortcomings. Newer ones are in development but may be expensive, and few studies have employed end-to-end video-based motor assessments. To fill this research gap, we developed a few-shot learning model based on a 3D residual and attention enhancement joint learning approach. We first collected video data from 48 HD patients and enhanced the key data. We then designed a 3D convolutional module based on channel-spatial attention to initially extract data features. After that, an excitation-based 3D residual convolution module was designed for feature learning. Finally, we used a fully connected layer based on channel-spatial attention to complete feature compression and obtain the assessment results. Our results show that the classification accuracy and recall of our model achieve state-of-the-art performance. The proposed lightweight HD disease motor assessment model is an improvement over other previous methods. Further developments of the methods are expected to provide a suitable baseline and follow-up assessment tool for future HD video-based studies. Furthermore, our proposed solution can be implemented with consumer-grade cameras, holding considerable potential for extensive implementation in HD telemedicine. The source code can be found at https://github.com/JackAILab/RAJNet.
22 April 2026 in Neurocomputing