Person image generation has become an increasingly important problem in computer vision with broad applications in virtual try-on, digital content creation, entertainment, and human-computer interaction. Despite recent advances in diffusion-based generative models that can produce photorealistic results, existing pipelines are still constrained by rigid modality requirements. Most prior methods rely on fixed patterns such as pose-plus-appearance inputs or text-only descriptions, limiting their flexibility and controllability in practical scenarios where users may prefer different or mixed types of guidance. This lack of adaptability poses a clear barrier to deployment in real-world systems. To overcome these challenges, we propose MaPPA (Multimodal Controllable Person Image Generation with Pose and Appearance Guidance), a unified multimodal framework that leverages transformer-based latent diffusion models. The central idea is to provide composable control over multiple modalities, enabling person image generation conditioned on text prompts, reference appearance images, pose keypoints, or any combination thereof. To achieve this, we introduce a unified framework that incorporates dedicated control blocks for appearance and pose guidance, which are strategically interleaved with transformer base blocks. These control blocks are conditioned on a unified multimodal embedding that integrates heterogeneous inputs into a consistent representation, thereby supporting arbitrary modality combinations with a single unified pipeline. Another key contribution of MaPPA is a cumulative classifier-free guidance strategy that enables allowing users to independently adjust the strength of appearance and pose guidance via scalar weights from different control signals. This design allows users to adjust the relative strength of appearance versus pose guidance, providing fine-grained controllability during inference. Furthermore, to address the common problem of detail loss in latent-diffusion decoders, we propose a texture enhancement decoding (TED) strategy, which fine-tunes the VAE decoder with edge-aware reconstruction objectives. This refinement significantly alleviates texture distortion, preserving high-frequency details in clothing patterns, facial regions, and other fine structures. Extensive experiments confirm that MaPPA achieves competitive quantitative scores while providing superior perceptual quality and user preference. Unlike task-specific methods, our framework-though constrained by the capacity of the unified multimodal embedding-supports combinations of text, pose, and appearance modalities within a single pipeline, demonstrating both flexibility and practical value. The code and the corresponding model will be made publicly available to the research community.

Understanding intention and emotion from social media poses unique challenges due to the inherent uncertainty in multimodal data, where posts often contain incomplete or missing modalities. While this uncertainty reflects real-world scenarios, it remains underexplored within the computer vision community, particularly in conjunction with the intrinsic relationship between emotion and intention. To address these challenges, we introduce the Multimodal IntentioN and Emotion Understanding in the Wild (MINE) dataset, comprising over 20,000 topic-specific social media posts with natural modality variations across text, image, video, and audio. MINE is distinctively constructed to capture both the uncertain nature of multimodal data and the implicit correlations between intentions and emotions, providing extensive annotations for both aspects. To tackle these scenarios, we propose the Bridging Emotion-Intention via Implicit Label Reasoning (BEAR) framework. BEAR consists of two key components: a BEIFormer that leverages emotion-intention correlations, and a Modality Asynchronous Prompt that handles modality uncertainty. Experiments show that BEAR outperforms existing methods in processing uncertain multimodal data while effectively mining emotion-intention relationships for social media content understanding.

Fashion image editing aims to edit an input image to obtain richer or distinct visual clothing matching effects. Existing global fashion image editing methods are difficult to achieve rich outfit combination effects while local fashion image editing is more in line with the needs of diverse and personalized outfit matching. The local editing techniques typically depend on text and auxiliary modalities (e.g., human poses, human keypoints, garment sketches, etc.) for image manipulation, where the auxiliary modalities essentially assist in locating the editing region. Since these auxiliary modalities usually involve additional efforts in practical application scenarios, text-driven fashion image editing shows high flexibility. In this paper, we propose TexFit, a Text-driven Fashion image Editing method using diffusion models, which performs the local image editing only with the easily accessible text. Our approach employs a text-based editing region location module to predict precise editing region in the fashion image. Then, we take the predicted region as the generation condition of diffusion models together with the text prompt to achieve precise local editing of fashion images while keeping the rest part intact. In addition, previous fashion datasets usually focus on global description, lacking local descriptive information that can guide the precise local editing. Therefore, we develop a new DFMM-Spotlight dataset by using region extraction and attribute combination strategies. It focuses locally on clothes and accessories, enabling local editing with text input. Experimental results on the DFMM-Spotlight dataset demonstrate the effectiveness of our model. Code and Datasets are available at https://texfit.github.io/.

High-dimensional, localized ribonucleic acid (RNA) sequencing is now possible owing to recent developments in spatial transcriptomics (ST). ST is based on highly multiplexed sequence analysis and uses barcodes to match the sequenced reads to their respective tissue locations. ST expression data suffer from high noise and dropout events; however, smoothing techniques have the promise to improve the data interpretability prior to performing downstream analyses. Single-cell RNA sequencing (scRNA-seq) data similarly suffer from these limitations, and smoothing methods developed for scRNA-seq can only utilize associations in transcriptome space (also known as one-factor smoothing methods). Since they do not account for spatial relationships, these one-factor smoothing methods cannot take full advantage of ST data. In this study, we present a novel two-factor smoothing technique, spatial and pattern combined smoothing (SPCS), that employs the k-nearest neighbor (kNN) technique to utilize information from transcriptome and spatial relationships. By performing SPCS on multiple ST slides from pancreatic ductal adenocarcinoma (PDAC), dorsolateral prefrontal cortex (DLPFC) and simulated high-grade serous ovarian cancer (HGSOC) datasets, smoothed ST slides have better separability, partition accuracy and biological interpretability than the ones smoothed by preexisting one-factor methods. Source code of SPCS is provided in Github (https://github.com/Usos/SPCS).

Domain Adaptation (DA) targets at adapting a model trained over the well-labeled source domain to the unlabeled target domain lying in different distributions. Existing DA normally assumes the well-labeled source domain is class-wise balanced, which means the size per source class is relatively similar. However, in real-world applications, labeled samples for some categories in the source domain could be extremely few due to the difficulty of data collection and annotation, which leads to decreasing performance over target domain on those few-shot categories. To perform fair cross-domain adaptation and boost the performance on these minority categories, we develop a novel Generative Few-shot Cross-domain Adaptation (GFCA) algorithm for fair cross-domain classification. Specifically, generative feature augmentation is explored to synthesize effective training data for few-shot source classes, while effective cross-domain alignment aims to adapt knowledge from source to facilitate the target learning. Experimental results on two large cross-domain visual datasets demonstrate the effectiveness of our proposed method on improving both few-shot and overall classification accuracy comparing with the state-of-the-art DA approaches.

Whole-Slide Histopathology Image (WSI) is generally considered the gold standard for cancer diagnosis and prognosis. Given the large inter-operator variation among pathologists, there is an imperative need to develop machine learning models based on WSIs for consistently predicting patient prognosis. The existing WSI-based prediction methods do not utilize the ordinal ranking loss to train the prognosis model, and thus cannot model the strong ordinal information among different patients in an efficient way. Another challenge is that a WSI is of large size (e.g., 100,000-by-100,000 pixels) with heterogeneous patterns but often only annotated with a single WSI-level label, which further complicates the training process. To address these challenges, we consider the ordinal characteristic of the survival process by adding a ranking-based regularization term on the Cox model and propose a weakly supervised deep ordinal Cox model (BDOCOX) for survival prediction from WSIs. Here, we generate amounts of bags from WSIs, and each bag is comprised of the image patches representing the heterogeneous patterns of WSIs, which is assumed to match the WSI-level labels for training the proposed model. The effectiveness of the proposed method is well validated by theoretical analysis as well as the prognosis and patient stratification results on three cancer datasets from The Cancer Genome Atlas (TCGA).

The integrative analysis of histopathological images and genomic data has received increasing attention for studying the complex mechanisms of driving cancers. However, most image-genomic studies have been restricted to combining histopathological images with the single modality of genomic data (e.g., mRNA transcription or genetic mutation), and thus neglect the fact that the molecular architecture of cancer is manifested at multiple levels, including genetic, epigenetic, transcriptional, and post-transcriptional events. To address this issue, we propose a novel ordinal multi-modal feature selection (OMMFS) framework that can simultaneously identify important features from both pathological images and multi-modal genomic data (i.e., mRNA transcription, copy number variation, and DNA methylation data) for the prognosis of cancer patients. Our model is based on a generalized sparse canonical correlation analysis framework, by which we also take advantage of the ordinal survival information among different patients for survival outcome prediction. We evaluate our method on three early-stage cancer datasets derived from The Cancer Genome Atlas (TCGA) project, and the experimental results demonstrated that both the selected image and multi-modal genomic markers are strongly correlated with survival enabling effective stratification of patients with distinct survival than the comparing methods, which is often difficult for early-stage cancer patients.

MOTIVATION: Rapid advances in single cell RNA sequencing (scRNA-seq) have produced higher-resolution cellular subtypes in multiple tissues and species. Methods are increasingly needed across datasets and species to (i) remove systematic biases, (ii) model multiple datasets with ambiguous labels and (iii) classify cells and map cell type labels. However, most methods only address one of these problems on broad cell types or simulated data using a single model type. It is also important to address higher-resolution cellular subtypes, subtype labels from multiple datasets, models trained on multiple datasets simultaneously and generalizability beyond a single model type. RESULTS: We developed a species- and dataset-independent transfer learning framework (LAmbDA) to train models on multiple datasets (even from different species) and applied our framework on simulated, pancreas and brain scRNA-seq experiments. These models mapped corresponding cell types between datasets with inconsistent cell subtype labels while simultaneously reducing batch effects. We achieved high accuracy in labeling cellular subtypes (weighted accuracy simulated 1 datasets: 90%; simulated 2 datasets: 94%; pancreas datasets: 88% and brain datasets: 66%) using LAmbDA Feedforward 1 Layer Neural Network with bagging. This method achieved higher weighted accuracy in labeling cellular subtypes than two other state-of-the-art methods, scmap and CaSTLe in brain (66% versus 60% and 32%). Furthermore, it achieved better performance in correctly predicting ambiguous cellular subtype labels across datasets in 88% of test cases compared with CaSTLe (63%), scmap (50%) and MetaNeighbor (50%). LAmbDA is model- and dataset-independent and generalizable to diverse data types representing an advance in biocomputing. AVAILABILITY AND IMPLEMENTATION: github.com/tsteelejohnson91/LAmbDA. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

BACKGROUND: Gene Co-expression Network Analysis (GCNA) helps identify gene modules with potential biological functions and has become a popular method in bioinformatics and biomedical research. However, most current GCNA algorithms use correlation to build gene co-expression networks and identify modules with highly correlated genes. There is a need to look beyond correlation and identify gene modules using other similarity measures for finding novel biologically meaningful modules. RESULTS: We propose a new generalized gene co-expression analysis algorithm via subspace clustering that can identify biologically meaningful gene co-expression modules with genes that are not all highly correlated. We use low-rank representation to construct gene co-expression networks and local maximal quasi-clique merger to identify gene co-expression modules. We applied our method on three large microarray datasets and a single-cell RNA sequencing dataset. We demonstrate that our method can identify gene modules with different biological functions than current GCNA methods and find gene modules with prognostic values. CONCLUSIONS: The presented method takes advantage of subspace clustering to generate gene co-expression networks rather than using correlation as the similarity measure between genes. Our generalized GCNA method can provide new insights from gene expression datasets and serve as a complement to current GCNA algorithms.