In this paper, we study the composed query image retrieval, which aims at retrieving the target image similar to the composed query, i.e., a reference image and the desired modification text. Compared with conventional image retrieval, this task is more challenging as it not only requires precisely aligning the composed query and target image in a common embedding space, but also simultaneously extracting related information from the reference image and modification text. In order to properly extract related information from the composed query, existing methods usually embed vision-language inputs using different feature encoders, e.g., CNN for images and LSTM/BERT for text, and then employ a complicated manually-designed composition module for learning the joint image-text representation. However, the architecture discrepancy in feature encoders would restrict the vision-language plenitudinous interaction. Meanwhile, certain complicated composition designs might significantly hamper the generalization ability of the model. To tackle these problems, we propose a new framework termed ComqueryFormer, which effectively processes the composed query with the Transformer for this task. Specifically, to eliminate the architecture discrepancy, we leverage a unified transformer-based architecture to homogeneously encode the vision-language inputs. Meanwhile, instead of the complicated composition module, the neat yet effective cross-modal transformer is adopted to hierarchically fuse the composed query at various vision scales. On the other hand, we introduce an efficient global-local alignment module to narrow the distance between the composed query and the target image. It not only considers the divergence in the global joint embedding space but also forces the model to focus on the local detail differences. Extensive experiments on three real-world datasets demonstrate the superiority of our ComqueryFormer. Our code can be found at: https://github.com/uestc-xyh/ComqueryFormer .

Despite its significant progress, cross-modal retrieval still suffers from one-to-many matching cases, where the multiplicity of semantic instances in another modality could be acquired by a given query. However, existing approaches usually map heterogeneous data into the learned space as deterministic point vectors. In spite of their remarkable performance in matching the most similar instance, such deterministic point embedding suffers from the insufficient representation of rich semantics in one-to-many correspondence. To address the limitations, we intuitively extend a deterministic point into a closed geometry and develop geometric representation learning methods for cross-modal retrieval. Thus, a set of points inside such a geometry could be semantically related to many candidates, and we could effectively capture the semantic uncertainty. We then introduce two types of geometric matching for one-to-many correspondence, i.e., point-to-rectangle matching (dubbed P2RM) and rectangle-to-rectangle matching (termed R2RM). The former treats all retrieved candidates as rectangles with zero volume (equivalent to points) and the query as a box, while the latter encodes all heterogeneous data into rectangles. Therefore, we could evaluate semantic similarity among heterogeneous data by the Euclidean distance from a point to a rectangle or the volume of intersection between two rectangles. Additionally, both strategies could be easily employed for off-the-self approaches and further improve the retrieval performance of baselines. Under various evaluation metrics, extensive experiments and ablation studies on several commonly used datasets, two for image-text matching and two for video-text retrieval, demonstrate our effectiveness and superiority.

Vein images generally appear darker with low contrast, which require contrast enhancement during preprocessing to design satisfactory hand vein recognition system. However, the modification introduced by contrast enhancement (CE) is reported to bring side effects through pixel intensity distribution adjustments. Furthermore, the inevitable results of fake vein generation or information loss occur and make nearly all vein recognition systems unconvinced. In this paper, a “CE-free” quality-specific vein recognition system is proposed, and three improvements are involved. First, a high-quality lab-vein capturing device is designed to solve the problem of low contrast from the view of hardware improvement. Then, a high quality lab-made database is established. Second, CFISH score, a fast and effective measurement for vein image quality evaluation, is proposed to obtain quality index of lab-made vein images. Then, unsupervised $K$ -means with optimized initialization and convergence condition is designed with the quality index to obtain the grouping results of the database, namely, low quality (LQ) and high quality (HQ). Finally, discriminative local binary pattern (DLBP) is adopted as the basis for feature extraction. For the HQ image, DLBP is adopted directly for feature extraction, and for the LQ one. CE_DLBP could be utilized for discriminative feature extraction for LQ images. Based on the lab-made database, rigorous experiments are conducted to demonstrate the effectiveness and feasibility of the proposed system. What is more, an additional experiment with PolyU database illustrates its generalization ability and robustness.

The accumulation of adversarial perturbations in the feature space makes it impossible for Deep Neural Networks (DNNs) to know what features are robust and reliable, and thus DNNs can be fooled by relying on a single contaminated feature. Numerous defense strategies attempt to improve their robustness by denoising, deactivating, or recalibrating non-robust features. Despite their effectiveness, we still argue that these methods are under-explored in terms of determining how trustworthy the features are. To address this issue, we propose a novel Evidence-based Multi-Feature Fusion (termed EMFF) for adversarial robustness. Specifically, our EMFF approach introduces evidential deep learning to help DNNs quantify the belief mass and uncertainty of the contaminated features. Subsequently, a novel multi-feature evidential fusion mechanism based on Dempster's rule is proposed to fuse the trusted features of multiple blocks within an architecture, which further helps DNNs avoid the induction of a single manipulated feature and thus improve their robustness. Comprehensive experiments confirm that compared with existing defense techniques, our novel EMFF method has obvious advantages and effectiveness in both scenarios of white-box and black-box attacks, and also prove that by integrating into several adversarial training strategies, we can improve the robustness of across distinct architectures, including traditional CNNs and recent vision Transformers with a few extra parameters and almost the same cost.

As an important visual understanding task, scene graph generation has been drawing widespread attention and could boost a broad range of downstream vision applications. Traditional scene graph generation methods based on different context refinements are trained with probabilistic chain rule, which treats objects and relationships as independent entities. Despite their surprisingly great progress, such a plain formulation unconsciously ignores the latent geometric structure of entities and relationships. To address this issue, we move beyond the traditional real-valued representations and use Quat ernion R elation E mbedding (QuatRE) to generate scene graphs with more expressive hypercomplex representations. More specifically, we introduce the concept of quaternion representations, hyper-complex valued with three imaginary components for objects entities, then formulate the relation triplets with Hamilton product. Benefiting from explicitly modeling the latent inter-dependencies among all imaginary components and strong expressive capacity, our proposed QuatRE method could better capture the interactions between entities. More importantly, our novel QuatRE method can be treated as a plug-in and well generalized into other methods for performance improvement as it involves no additional layers. Finally, extensive comparisons of our proposed method against the state-of-the-art methods on two large-scale and widely-used datasets, i.e. Visual Genome and Open Images, demonstrated our superiority and generalization capability on various metrics for biased or unbiased inference.

Despite being highly secure, vein recognition suffers from high inter-class similarity and intra-class variation resulting from uncontrolled image capture, making the design of discriminative and robust representation very important. The recent success of convolutional neural network (CNN) for various image understanding tasks makes it a promising method for feature extraction. However, limited variability in small-scale datasets leads to systems derived from direct training or fine-tuning not transferable and unreliable for practical biometric applications. This motivates the design a multi-weighted co-occurrence descriptor encoding (MWCDE) model for vein recognition. Instead of directly conducting a feed-forward operation with a pre-trained CNN for obtaining the semantic features from the fully-connected layers, co-occurrence features among convolutional filters are modelled firstly in MWCDE by a simple convolution between an indicator filter in a higher layer with a to-be-reweighted filter in a lower layer, and a redundancy-driven indicator filter selection algorithm is designed for filtering out some ambiguous representations. Secondly, another hard feature weighting strategy with a binary masking scheme is proposed for discarding noisy background and feature redundancy. The selected high-order descriptors are then embedded and aggregated into compact feature vectors with a saliency driven spatial weighted Fisher vector algorithm, followed by the introduction of a generalized support vector machine for recognition. Extensive experiments with three benchmark vein datasets demonstrate that the proposed framework can achieve state-of-the-art results, and an additional experiment with the PolyU multispectral palmprint database illustrates its generalization ability. Code will be available at 送https://github.com/RobinCSIRO/MWCDE-for-Vein-Recognitioni适.