Hash center-based deep hashing methods improve upon pairwise or triplet-based approaches by assigning fixed hash centers to each class as learning targets, thereby avoiding the inefficiency of local similarity optimization. However, random center initialization often disregards inter-class semantic relationships. While existing two-stage methods mitigate this by first refining hash centers with semantics and then training the hash function, they introduce additional complexity, computational overhead, and suboptimal performance due to stage-wise discrepancies. To address these limitations, we propose Center-Reassigned Hashing (CRH), an end-to-end framework that dynamically reassigns hash centers from a preset codebook while jointly optimizing the hash function. Unlike previous methods, CRH adapts hash centers to the data distribution without explicit center optimization phases, enabling seamless integration of semantic relationships into the learning process. Furthermore, a multi-head mechanism enhances the representational capacity of hash centers, capturing richer semantic structures. Extensive experiments on three benchmarks demonstrate that CRH learns semantically meaningful hash centers and outperforms state-of-the-art deep hashing methods in retrieval tasks.

Thermal runaway of the power battery can pose serious security risks. Therefore, it is imperative to monitor and control battery temperature in real time. To address the shortcomings of existing fiber Bragg grating (FBG) based temperature demodulation methods on efficiency, accuracy, and cost, a chirped-FBG (CFBG) slope-assisted weak-FBG (wFBG) demodulation method is proposed in this paper. The core mechanism of this method involves using the rising edge of the CFBG transmission spectrum as a wavelength-selective filter for wavelength-addressing series-connected identical wFBGs. In the proposed system, all optical components are fiber-coupled, thereby providing advantages such as low cost and compact size. The experimental results reveal that the system can operate stably at a demodulation rate of 1 Hz within 33°C to 107°C, with a demodulation standard error of 0.0831°C. The temperature demodulation resolution is 1.73×10-4°C, equivalent to a spectrometer demodulation resolution of 1.6 fm. By deploying the system on battery cells, online monitoring of global temperature in power batteries with high sensing density can be realized, providing a feasible technical solution to enhance monitoring precision, real-time performance, economic efficiency, and implementation convenience for the temperature monitoring requirements of battery management systems and thermal management systems.

Bioluminescence tomography (BLT) can provide non-invasive quantitative three-dimensional tumor information which has been widely applied in pre-clinical studies. Meanwhile, in recent years, deep learning methods have significantly improved the reconstruction resolution and speed by establishing a non-linear mapping relationship between surface-measured bioluminescence and light source distribution. However, this mapping relationship only works for specific biological tissues and light transmission processes under fixed wavelengths, resulting in poor stability and generalizability. To meet the requirements of diverse practical scenarios and inspired by more effective sparse regularization and graph representation theory, we propose a novel Graph Attention Iterative Contraction Network (GAICN) to conduct a finite element mesh spatial representation study. In the GAICN framework, two learnable spatial topological transforms based on the graph attention mechanism and an iterative contraction activation function were devised to achieve non-local feature aggregation and dynamic adjustment of weights between first-order neighboring nodes in the mesh. As a deep unrolling method, GAICN naturally inherits the coherence of surface bioluminescence with the light source in Forward-Backward Splitting (FBS), thus enhancing the generalizability, stability and interpretability of the network. Both simulation and in-vivo experiments further indicated that GAICN achieved superior reconstruction performance in terms of spatial location, dual light source resolution, stability, generalizability, as well as in-vivo practicability.

Breast cancer, the predominant malignancy among women, is characterized by significant heterogeneity, leading to the emergence of distinct molecular subtypes. Accurate differentiation of these molecular subtypes holds paramount clinical significance, owing to substantial variations in prognosis, therapeutic strategies, and survival outcomes. In this study, we propose a cross-sequence joint representation and hypergraph convolution network (CORONet) for classifying molecular subtypes of breast cancer using incomplete DCE-MRI. Specifically, we first build a cross-sequence joint representation (COR) module to integrate image imputation and feature representation into a unified framework, encouraging effective feature extraction for subsequent classification. Then, we fuse multiple COR features and applied feature selection to reduce the redundant information between sequences. Finally, we deploy hypergraph structures to model high-order correlation among different subjects and extracted high-level semantic features by hypergraph convolutions for molecular subtyping. Extensive experiments on incomplete DCE-MRIs of 395 patients from the TCIA repository showed a significant improvement of our CORONet over state of the arts, with the area under the curve (AUC) of 0.891 and 0.903 for luminal and triple-negative (TN) subtype prediction, respectively. Similar advantages of CORONet were also confirmed in partial complete DCE-MRIs of 144 patients, achieving an AUC of 0.858 and 0.832 for predicting luminal and TN subtypes of breast cancer, respectively. Nevertheless, both of these values were lower compared to the scenario where DCE-MRIs from all 395 patients were utilized. Our study contributes to the precise molecular subtyping using incomplete multi-sequence DCE-MRI, thereby offering promising prospects for future risk stratification of breast cancer patients.

Orbital lymphoproliferative disorders (OLPDs) are complex orbital mass-like lesions ranging from benign to malignant. Precise preoperative diagnosis of OLPDs holds profound importance in facilitating timely and effective patient management. Recent studies have shown that exploiting multimodal images can boost the performance in identifying different orbital lesions. However, one or several imaging modalities are sometimes missing in practical applications, which has not yet been properly addressed in existing studies. To this end, we propose a novel OLPD diagnostic method with incomplete multimodal images based on self-/cross-representation and hypergraph ensemble. Specifically, in the first stage, we develop a self-representation network to extract unimodal features and a cross-representation network to impute missing features. In the second stage, by using unimodal features as input, we construct a hypergraph for each modality to make unimodal diagnosis; while for multimodal diagnosis we conduct a multi-view grouping fusion method to reduce the semantic gap between multimodal features and fuse multiple unimodal hypergraphs as multimodal hypergraph to perform multimodal diagnosis. In the third stage, we propose an ensemble strategy that incorporates unimodal diagnosis and multimodal diagnosis to accomplish the final decision. Extensive experiments demonstrate that the proposed model outperforms the state-of-the-art approaches.

The diversity of recommendation has attracted a lot of attention in recommender systems due to its ability to improve user experience. Most of the diversified recommendation tasks usually exploit user-item interaction records for mining user explicit preferences, while rarely explore the user-item non-interaction records. For diversified recommendations, however, the neglected non-interaction records are especially important for capturing users’ potential interests to improve the diversity of recommendation. Moreover, the majority of diversified recommendation methods run in two stages: first optimizing the users and items embeddings by relevance, then generating the diversified items list by post-processing methods. These methods are not end-to-end thus can hardly reach global optimum. To solve above limitations, we propose an end-to-end Dynamic Diversified Graph framework (DDGraph) which constructs the user-item graph dynamically based on the users and items embeddings. Technically, we initialize a user-item interaction graph and dynamically update the graph by selecting a set of diverse items for each user and building links between the items and user. The selection of diverse items can be achieved by different candidate selection operators. Specifically, we design a Quantile Progressive Candidate Selection (QPCS) operator based on the latent space division. To the best of our knowledge, our method is the first to diversify recommendation results by dynamic end-to-end graph construction and the QPCS has a higher computational efficiency than other operators. Extensive experiments on the benchmark dataset illustrate the effectiveness and superiority of the DDGraph framework.

Hashing, which represents data items as compact binary codes, has been becoming a more and more popular technique, e.g., for large-scale image retrieval, owing to its super fast search speed as well as its extremely economical memory consumption. However, existing hashing methods all try to learn binary codes from artificially balanced datasets which are not commonly available in real-world scenarios. In this paper, we propose Long-Tail Hashing Network (LTHNet), a novel two-stage deep hashing approach that addresses the problem of learning to hash for more realistic datasets where the data labels roughly exhibit a long-tail distribution. Specifically, the first stage is to learn relaxed embeddings of the given dataset with its long-tail characteristic taken into account via an end-to-end deep neural network; the second stage is to binarize those obtained embeddings. A critical part of LTHNet is its dynamic meta-embedding module extended with a determinantal point process which can adaptively realize visual knowledge transfer between head and tail classes, and thus enrich image representations for hashing. Our experiments have shown that LTHNet achieves dramatic performance improvements over all state-of-the-art competitors on long-tail datasets, with no or little sacrifice on balanced datasets. Further analyses reveal that while to our surprise directly manipulating class weights in the loss function has little effect, the extended dynamic meta-embedding module, the usage of cross-entropy loss instead of square loss, and the relatively small batch-size for training all contribute to LTHNet's success.

Instance segmentation and panoptic segmentation is being paid more and more attention in recent years. In comparison with bounding box based object detection and semantic segmentation, instance segmentation can provide more analytical results at pixel level. Given the insight that pixels belonging to one instance have one or more common attributes of current instance, we bring up an one-stage instance segmentation network named Common Attribute Support Network (CASNet), which realizes instance segmentation by predicting and clustering common attributes. CASNet is designed in the manner of fully convolutional and can implement training and inference from end to end. And CASNet manages predicting the instance without overlaps and holes, which problem exists in most of current instance segmentation algorithms. Furthermore, it can be easily extended to panoptic segmentation through minor modifications with little computation overhead. CASNet builds a bridge between semantic and instance segmentation from finding pixel class ID to obtaining class and instance ID by operations on common attribute. Through experiment for instance and panoptic segmentation, CASNet gets mAP 32.8% and PQ 59.0% on Cityscapes validation dataset by joint training, and mAP 36.3% and PQ 66.1% by separated training mode. For panoptic segmentation, CASNet gets state-of-the-art performance on the Cityscapes validation dataset.

Annotating images with tags is useful for indexing and retrieving images. However, many available annotation data include missing or inaccurate annotations. In this paper, we propose an image annotation framework which sequentially performs tag completion and refinement. We utilize the subspace property of data via sparse subspace clustering for tag completion. Then we propose a novel matrix completion model for tag refinement, integrating visual correlation, semantic correlation and the novelly studied property of complex errors. The proposed method outperforms the state-of-the-art approaches on multiple benchmark datasets even when they contain certain levels of annotation noise.

Tag-based image retrieval (TBIR) has drawn much attention in recent years due to the explosive amount of digital images and crowdsourcing tags. However, the TBIR applications still suffer from the deficient and inaccurate tags provided by users. Inspired by the subspace clustering methods, we formulate the tag completion problem in a subspace clustering model which assumes that images are sampled from subspaces, and complete the tags using the state-of-the-art Low Rank Representation (LRR) method. And we propose a matrix completion algorithm to further refine the tags. Our empirical results on multiple benchmark datasets for image annotation show that the proposed algorithm outperforms state-of-the-art approaches when handling missing and noisy tags.