Search

In recent years, Multi-Modality Image Fusion (MMIF) has been applied to many fields, which has attracted many scholars to endeavour to improve the fusion performance. However, the prevailing focus has predominantly been on the architecture design, rather than the training strategies. As a low-level vision task, image fusion is supposed to quickly deliver output images for observation and supporting downstream tasks. Thus, superfluous computational and storage overheads should be avoided. In this work, a lightweight Distilled Mini-Model with a Dynamic Refresh strategy (MMDRFuse) is proposed to achieve this objective. To pursue model parsimony, an extremely small convolutional network with a total of 113 trainable parameters (0.44 KB) is obtained by three carefully designed supervisions. First, digestible distillation is constructed by emphasising external spatial feature consistency, delivering soft supervision with balanced details and saliency for the target network. Second, we develop a comprehensive loss to balance the pixel, gradient, and perception clues from the source images. Third, an innovative dynamic refresh training strategy is used to collaborate history parameters and current supervision during training, together with an adaptive adjust function to optimise the fusion network. Extensive experiments on several public datasets demonstrate that our method exhibits promising advantages in terms of model efficiency and complexity, with superior performance in multiple image fusion tasks and downstream pedestrian detection application. The code of this work is publicly available at https://github.com/yanglinDeng/MMDRFuse., Comment: 10 pages, 8 figures, accpeted by ACM International Conference on Multimedia 2024(Oral)

Compositional actions consist of dynamic (verbs) and static (objects) concepts. Humans can easily recognize unseen compositions using the learned concepts. For machines, solving such a problem requires a model to recognize unseen actions composed of previously observed verbs and objects, thus requiring so-called compositional generalization ability. To facilitate this research, we propose a novel Zero-Shot Compositional Action Recognition (ZS-CAR) task. For evaluating the task, we construct a new benchmark, Something-composition (Sth-com), based on the widely used Something-Something V2 dataset. We also propose a novel Component-to-Composition (C2C) learning method to solve the new ZS-CAR task. C2C includes an independent component learning module and a composition inference module. Last, we devise an enhanced training strategy to address the challenges of component variations between seen and unseen compositions and to handle the subtle balance between learning seen and unseen actions. The experimental results demonstrate that the proposed framework significantly surpasses the existing compositional generalization methods and sets a new state-of-the-art. The new Sth-com benchmark and code are available at https://github.com/RongchangLi/ZSCAR_C2C., Comment: Accepted by ECCV2024

Vision transformers combined with self-supervised learning have enabled the development of models which scale across large datasets for several downstream tasks like classification, segmentation and detection. The low-shot learning capability of these models, across several low-shot downstream tasks, has been largely under explored. We perform a system level study of different self supervised pretext tasks, namely contrastive learning, clustering, and masked image modelling for their low-shot capabilities by comparing the pretrained models. In addition we also study the effects of collapse avoidance methods, namely centring, ME-MAX, sinkhorn, on these downstream tasks. Based on our detailed analysis, we introduce a framework involving both mask image modelling and clustering as pretext tasks, which performs better across all low-shot downstream tasks, including multi-class classification, multi-label classification and semantic segmentation. Furthermore, when testing the model on full scale datasets, we show performance gains in multi-class classification, multi-label classification and semantic segmentation.

Masked Image Modeling (MIM)-based models, such as SdAE, CAE, GreenMIM, and MixAE, have explored different strategies to enhance the performance of Masked Autoencoders (MAE) by modifying prediction, loss functions, or incorporating additional architectural components. In this paper, we propose an enhanced approach that boosts MAE performance by integrating pseudo labelling for both class and data tokens, alongside replacing the traditional pixel-level reconstruction with token-level reconstruction. This strategy uses cluster assignments as pseudo labels to promote instance-level discrimination within the network, while token reconstruction requires generation of discrete tokens encapturing local context. The targets for pseudo labelling and reconstruction needs to be generated by a teacher network. To disentangle the generation of target pseudo labels and the reconstruction of the token features, we decouple the teacher into two distinct models, where one serves as a labelling teacher and the other as a reconstruction teacher. This separation proves empirically superior to a single teacher, while having negligible impact on throughput and memory consumption. Incorporating pseudo-labelling as an auxiliary task has demonstrated notable improvements in ImageNet-1K and other downstream tasks, including classification, semantic segmentation, and detection.

There is currently strong interest in improving visual object tracking by augmenting the RGB modality with the output of a visual event camera that is particularly informative about the scene motion. However, existing approaches perform event feature extraction for RGB-E tracking using traditional appearance models, which have been optimised for RGB only tracking, without adapting it for the intrinsic characteristics of the event data. To address this problem, we propose an Event backbone (Pooler), designed to obtain a high-quality feature representation that is cognisant of the innate characteristics of the event data, namely its sparsity. In particular, Multi-Scale Pooling is introduced to capture all the motion feature trends within event data through the utilisation of diverse pooling kernel sizes. The association between the derived RGB and event representations is established by an innovative module performing adaptive Mutually Guided Fusion (MGF). Extensive experimental results show that our method significantly outperforms state-of-the-art trackers on two widely used RGB-E tracking datasets, including VisEvent and COESOT, where the precision and success rates on COESOT are improved by 4.9% and 5.2%, respectively. Our code will be available at https://github.com/SSSpc333/TENet.

RGBT tracking draws increasing attention due to its robustness in multi-modality warranting (MMW) scenarios, such as nighttime and bad weather, where relying on a single sensing modality fails to ensure stable tracking results. However, the existing benchmarks predominantly consist of videos collected in common scenarios where both RGB and thermal infrared (TIR) information are of sufficient quality. This makes the data unrepresentative of severe imaging conditions, leading to tracking failures in MMW scenarios. To bridge this gap, we present a new benchmark, MV-RGBT, captured specifically in MMW scenarios. In contrast with the existing datasets, MV-RGBT comprises more object categories and scenes, providing a diverse and challenging benchmark. Furthermore, for severe imaging conditions of MMW scenarios, a new problem is posed, namely \textit{when to fuse}, to stimulate the development of fusion strategies for such data. We propose a new method based on a mixture of experts, namely MoETrack, as a baseline fusion strategy. In MoETrack, each expert generates independent tracking results along with the corresponding confidence score, which is used to control the fusion process. Extensive experimental results demonstrate the significant potential of MV-RGBT in advancing RGBT tracking and elicit the conclusion that fusion is not always beneficial, especially in MMW scenarios. Significantly, the proposed MoETrack method achieves new state-of-the-art results not only on MV-RGBT, but also on standard benchmarks, such as RGBT234, LasHeR, and the short-term split of VTUAV (VTUAV-ST). More information of MV-RGBT and the source code of MoETrack will be released at https://github.com/Zhangyong-Tang/MoETrack.

Pooling is a crucial operation in computer vision, yet the unique structure of skeletons hinders the application of existing pooling strategies to skeleton graph modelling. In this paper, we propose an Improved Graph Pooling Network, referred to as IGPN. The main innovations include: Our method incorporates a region-awareness pooling strategy based on structural partitioning. The correlation matrix of the original feature is used to adaptively adjust the weight of information in different regions of the newly generated features, resulting in more flexible and effective processing. To prevent the irreversible loss of discriminative information, we propose a cross fusion module and an information supplement module to provide block-level and input-level information respectively. As a plug-and-play structure, the proposed operation can be seamlessly combined with existing GCN-based models. We conducted extensive evaluations on several challenging benchmarks, and the experimental results indicate the effectiveness of our proposed solutions. For example, in the cross-subject evaluation of the NTU-RGB+D 60 dataset, IGPN achieves a significant improvement in accuracy compared to the baseline while reducing Flops by nearly 70%; a heavier version has also been introduced to further boost accuracy.

Recently, masked image modeling (MIM), an important self-supervised learning (SSL) method, has drawn attention for its effectiveness in learning data representation from unlabeled data. Numerous studies underscore the advantages of MIM, highlighting how models pretrained on extensive datasets can enhance the performance of downstream tasks. However, the high computational demands of pretraining pose significant challenges, particularly within academic environments, thereby impeding the SSL research progress. In this study, we propose efficient training recipes for MIM based SSL that focuses on mitigating data loading bottlenecks and employing progressive training techniques and other tricks to closely maintain pretraining performance. Our library enables the training of a MAE-Base/16 model on the ImageNet 1K dataset for 800 epochs within just 18 hours, using a single machine equipped with 8 A100 GPUs. By achieving speed gains of up to 5.8 times, this work not only demonstrates the feasibility of conducting high-efficiency SSL training but also paves the way for broader accessibility and promotes advancement in SSL research particularly for prototyping and initial testing of SSL ideas. The code is available in https://github.com/erow/FastSSL.

In a scenario where multi-modal cameras are operating together, the problem of working with non-aligned images cannot be avoided. Yet, existing image fusion algorithms rely heavily on strictly registered input image pairs to produce more precise fusion results, as a way to improve the performance of downstream high-level vision tasks. In order to relax this assumption, one can attempt to register images first. However, the existing methods for registering multiple modalities have limitations, such as complex structures and reliance on significant semantic information. This paper aims to address the problem of image registration and fusion in a single framework, called BusRef. We focus on Infrared-Visible image registration and fusion task (IVRF). In this framework, the input unaligned image pairs will pass through three stages: Coarse registration, Fine registration and Fusion. It will be shown that the unified approach enables more robust IVRF. We also propose a novel training and evaluation strategy, involving the use of masks to reduce the influence of non-reconstructible regions on the loss functions, which greatly improves the accuracy and robustness of the fusion task. Last but not least, a gradient-aware fusion network is designed to preserve the complementary information. The advanced performance of this algorithm is demonstrated by

Advanced image fusion methods are devoted to generating the fusion results by aggregating the complementary information conveyed by the source images. However, the difference in the source-specific manifestation of the imaged scene content makes it difficult to design a robust and controllable fusion process. We argue that this issue can be alleviated with the help of higher-level semantics, conveyed by the text modality, which should enable us to generate fused images for different purposes, such as visualisation and downstream tasks, in a controllable way. This is achieved by exploiting a vision-and-language model to build a coarse-to-fine association mechanism between the text and image signals. With the guidance of the association maps, an affine fusion unit is embedded in the transformer network to fuse the text and vision modalities at the feature level. As another ingredient of this work, we propose the use of textual attention to adapt image quality assessment to the fusion task. To facilitate the implementation of the proposed text-guided fusion paradigm, and its adoption by the wider research community, we release a text-annotated image fusion dataset IVT. Extensive experiments demonstrate that our approach (TextFusion) consistently outperforms traditional appearance-based fusion methods. Our code and dataset will be publicly available at https://github.com/AWCXV/TextFusion., Comment: v2 version, 13 pages, 16 figures, with the code repository link

Recently, self-supervised metric learning has raised attention for the potential to learn a generic distance function. It overcomes the limitations of conventional supervised one, e.g., scalability and label biases. Despite progress in this domain, current benchmarks, incorporating a narrow scope of classes, stop the nuanced evaluation of semantic representations. To bridge this gap, we introduce a large-scale benchmark with diversity and granularity of classes, Statistical Metric Learning Benchmark (SMLB) built upon ImageNet-21K and WordNet. SMLB is designed to rigorously evaluate the discriminative discernment and generalizability across more than 14M images, 20K classes, and 16K taxonomic nodes. Alongside, we propose novel evaluation metrics -- `overlap' for separability and `aSTD' for consistency -- to measure distance statistical information, which are efficient and robust to the change of class number. Our benchmark offers a novel perspective of evaluating the quality of representations beyond accuracy. Our findings reveal the limitations of supervised learning and the class bias inherent in SSL models, offering insights into potential areas for future model enhancement.

Contrastive learning has achieved great success in skeleton-based action recognition. However, most existing approaches encode the skeleton sequences as entangled spatiotemporal representations and confine the contrasts to the same level of representation. Instead, this paper introduces a novel contrastive learning framework, namely Spatiotemporal Clues Disentanglement Network (SCD-Net). Specifically, we integrate the decoupling module with a feature extractor to derive explicit clues from spatial and temporal domains respectively. As for the training of SCD-Net, with a constructed global anchor, we encourage the interaction between the anchor and extracted clues. Further, we propose a new masking strategy with structural constraints to strengthen the contextual associations, leveraging the latest development from masked image modelling into the proposed SCD-Net. We conduct extensive evaluations on the NTU-RGB+D (60&120) and PKU-MMD (I&II) datasets, covering various downstream tasks such as action recognition, action retrieval, transfer learning, and semi-supervised learning. The experimental results demonstrate the effectiveness of our method, which outperforms the existing state-of-the-art (SOTA) approaches significantly.

Generative models (GMs) have received increasing research interest for their remarkable capacity to achieve comprehensive understanding. However, their potential application in the domain of multi-modal tracking has remained relatively unexplored. In this context, we seek to uncover the potential of harnessing generative techniques to address the critical challenge, information fusion, in multi-modal tracking. In this paper, we delve into two prominent GM techniques, namely, Conditional Generative Adversarial Networks (CGANs) and Diffusion Models (DMs). Different from the standard fusion process where the features from each modality are directly fed into the fusion block, we condition these multi-modal features with random noise in the GM framework, effectively transforming the original training samples into harder instances. This design excels at extracting discriminative clues from the features, enhancing the ultimate tracking performance. To quantitatively gauge the effectiveness of our approach, we conduct extensive experiments across two multi-modal tracking tasks, three baseline methods, and three challenging benchmarks. The experimental results demonstrate that the proposed generative-based fusion mechanism achieves state-of-the-art performance, setting new records on LasHeR and RGBD1K., Comment: 10 figures, 8 tables

Self-supervised pretraining (SSP) has emerged as a popular technique in machine learning, enabling the extraction of meaningful feature representations without labelled data. In the realm of computer vision, pretrained vision transformers (ViTs) have played a pivotal role in advancing transfer learning. Nonetheless, the escalating cost of finetuning these large models has posed a challenge due to the explosion of model size. This study endeavours to evaluate the effectiveness of pure self-supervised learning (SSL) techniques in computer vision tasks, obviating the need for finetuning, with the intention of emulating human-like capabilities in generalisation and recognition of unseen objects. To this end, we propose an evaluation protocol for zero-shot segmentation based on a prompting patch. Given a point on the target object as a prompt, the algorithm calculates the similarity map between the selected patch and other patches, upon that, a simple thresholding is applied to segment the target. Another evaluation is intra-object and inter-object similarity to gauge discriminatory ability of SSP ViTs. Insights from zero-shot segmentation from prompting and discriminatory abilities of SSP led to the design of a simple SSP approach, termed MMC. This approaches combines Masked image modelling for encouraging similarity of local features, Momentum based self-distillation for transferring semantics from global to local features, and global Contrast for promoting semantics of global features, to enhance discriminative representations of SSP ViTs. Consequently, our proposed method significantly reduces the overlap of intra-object and inter-object similarities, thereby facilitating effective object segmentation within an image. Our experiments reveal that MMC delivers top-tier results in zero-shot semantic segmentation across various datasets.

In recent years, numerous ideas have emerged for designing a mutually reinforcing mechanism or extra stages for the image fusion task, ignoring the inevitable gaps between different vision tasks and the computational burden. We argue that there is a scope to improve the fusion performance with the help of the FusionBooster, a model specifically designed for the fusion task. In particular, our booster is based on the divide-and-conquer strategy controlled by an information probe. The booster is composed of three building blocks: the probe units, the booster layer, and the assembling module. Given the result produced by a backbone method, the probe units assess the fused image and divide the results according to their information content. This is instrumental in identifying missing information, as a step to its recovery. The recovery of the degraded components along with the fusion guidance are the role of the booster layer. Lastly, the assembling module is responsible for piecing these advanced components together to deliver the output. We use concise reconstruction loss functions in conjunction with lightweight autoencoder models to formulate the learning task, with marginal computational complexity increase. The experimental results obtained in various fusion tasks, as well as downstream detection tasks, consistently demonstrate that the proposed FusionBooster significantly improves the performance. Our code will be publicly available at https://github.com/AWCXV/FusionBooster., Comment: 18 pages; v2, including the code repository

Deep learning based fusion methods have been achieving promising performance in image fusion tasks. This is attributed to the network architecture that plays a very important role in the fusion process. However, in general, it is hard to specify a good fusion architecture, and consequently, the design of fusion networks is still a black art, rather than science. To address this problem, we formulate the fusion task mathematically, and establish a connection between its optimal solution and the network architecture that can implement it. This approach leads to a novel method proposed in the paper of constructing a lightweight fusion network. It avoids the time-consuming empirical network design by a trial-and-test strategy. In particular we adopt a learnable representation approach to the fusion task, in which the construction of the fusion network architecture is guided by the optimisation algorithm producing the learnable model. The low-rank representation (LRR) objective is the foundation of our learnable model. The matrix multiplications, which are at the heart of the solution are transformed into convolutional operations, and the iterative process of optimisation is replaced by a special feed-forward network. Based on this novel network architecture, an end-to-end lightweight fusion network is constructed to fuse infrared and visible light images. Its successful training is facilitated by a detail-to-semantic information loss function proposed to preserve the image details and to enhance the salient features of the source images. Our experiments show that the proposed fusion network exhibits better fusion performance than the state-of-the-art fusion methods on public datasets. Interestingly, our network requires a fewer training parameters than other existing methods. The codes are available at https://github.com/hli1221/imagefusion-LRRNet, Comment: 14 pages, 15 figures, 8 tables

Transformers, which were originally developed for natural language processing, have recently generated significant interest in the computer vision and audio communities due to their flexibility in learning long-range relationships. Constrained by the data hungry nature of transformers and the limited amount of labelled data, most transformer-based models for audio tasks are finetuned from ImageNet pretrained models, despite the huge gap between the domain of natural images and audio. This has motivated the research in self-supervised pretraining of audio transformers, which reduces the dependency on large amounts of labeled data and focuses on extracting concise representations of audio spectrograms. In this paper, we propose \textbf{L}ocal-\textbf{G}lobal \textbf{A}udio \textbf{S}pectrogram v\textbf{I}sion \textbf{T}ransformer, namely ASiT, a novel self-supervised learning framework that captures local and global contextual information by employing group masked model learning and self-distillation. We evaluate our pretrained models on both audio and speech classification tasks, including audio event classification, keyword spotting, and speaker identification. We further conduct comprehensive ablation studies, including evaluations of different pretraining strategies. The proposed ASiT framework significantly boosts the performance on all tasks and sets a new state-of-the-art performance in five audio and speech classification tasks, outperforming recent methods, including the approaches that use additional datasets for pretraining.

RGB-D object tracking has attracted considerable attention recently, achieving promising performance thanks to the symbiosis between visual and depth channels. However, given a limited amount of annotated RGB-D tracking data, most state-of-the-art RGB-D trackers are simple extensions of high-performance RGB-only trackers, without fully exploiting the underlying potential of the depth channel in the offline training stage. To address the dataset deficiency issue, a new RGB-D dataset named RGBD1K is released in this paper. The RGBD1K contains 1,050 sequences with about 2.5M frames in total. To demonstrate the benefits of training on a larger RGB-D data set in general, and RGBD1K in particular, we develop a transformer-based RGB-D tracker, named SPT, as a baseline for future visual object tracking studies using the new dataset. The results, of extensive experiments using the SPT tracker emonstrate the potential of the RGBD1K dataset to improve the performance of RGB-D tracking, inspiring future developments of effective tracker designs. The dataset and codes will be available on the project homepage: https://github.com/xuefeng-zhu5/RGBD1K.

As a structured prediction task, scene graph generation, given an input image, aims to explicitly model objects and their relationships by constructing a visually-grounded scene graph. In the current literature, such task is universally solved via a message passing neural network based mean field variational Bayesian methodology. The classical loose evidence lower bound is generally chosen as the variational inference objective, which could induce oversimplified variational approximation and thus underestimate the underlying complex posterior. In this paper, we propose a novel doubly reparameterized importance weighted structure learning method, which employs a tighter importance weighted lower bound as the variational inference objective. It is computed from multiple samples drawn from a reparameterizable Gumbel-Softmax sampler and the resulting constrained variational inference task is solved by a generic entropic mirror descent algorithm. The resulting doubly reparameterized gradient estimator reduces the variance of the corresponding derivatives with a beneficial impact on learning. The proposed method achieves the state-of-the-art performance on various popular scene graph generation benchmarks., Comment: arXiv admin note: substantial text overlap with arXiv:2205.07017

Image set-based visual classification methods have achieved remarkable performance, via characterising the image set in terms of a non-singular covariance matrix on a symmetric positive definite (SPD) manifold. To adapt to complicated visual scenarios better, several Riemannian networks (RiemNets) for SPD matrix nonlinear processing have recently been studied. However, it is pertinent to ask, whether greater accuracy gains can be achieved by simply increasing the depth of RiemNets. The answer appears to be negative, as deeper RiemNets tend to lose generalization ability. To explore a possible solution to this issue, we propose a new architecture for SPD matrix learning. Specifically, to enrich the deep representations, we adopt SPDNet [1] as the backbone, with a stacked Riemannian autoencoder (SRAE) built on the tail. The associated reconstruction error term can make the embedding functions of both SRAE and of each RAE an approximate identity mapping, which helps to prevent the degradation of statistical information. We then insert several residual-like blocks with shortcut connections to augment the representational capacity of SRAE, and to simplify the training of a deeper network. The experimental evidence demonstrates that our DreamNet can achieve improved accuracy with increased depth of the network., Comment: 9 pages, 7 figures

Vision transformers have generated significant interest in the computer vision community because of their flexibility in exploiting contextual information, whether it is sharply confined local, or long range global. However, they are known to be data hungry. This has motivated the research in self-supervised transformer pretraining, which does not need to decode the semantic information conveyed by labels to link it to the image properties, but rather focuses directly on extracting a concise representation of the image data that reflects the notion of similarity, and is invariant to nuisance factors. The key vehicle for the self-learning process used by the majority of self-learning methods is the generation of multiple views of the training data and the creation of pretext tasks which use these views to define the notion of image similarity, and data integrity. However, this approach lacks the natural propensity to extract contextual information. We propose group masked model learning (GMML), a self-supervised learning (SSL) mechanism for pretraining vision transformers with the ability to extract the contextual information present in all the concepts in an image. GMML achieves this by manipulating randomly groups of connected tokens, ensuingly covering a meaningful part of a semantic concept, and then recovering the hidden semantic information from the visible part of the concept. GMML implicitly introduces a novel data augmentation process. Unlike most of the existing SSL approaches, GMML does not require momentum encoder, nor rely on careful implementation details such as large batches and gradient stopping, which are all artefacts of most of the current self-supervised learning techniques. The source code is publicly available for the community to train on bigger corpora: https://github.com/Sara-Ahmed/GMML.

Scene graph generation is a structured prediction task aiming to explicitly model objects and their relationships via constructing a visually-grounded scene graph for an input image. Currently, the message passing neural network based mean field variational Bayesian methodology is the ubiquitous solution for such a task, in which the variational inference objective is often assumed to be the classical evidence lower bound. However, the variational approximation inferred from such loose objective generally underestimates the underlying posterior, which often leads to inferior generation performance. In this paper, we propose a novel importance weighted structure learning method aiming to approximate the underlying log-partition function with a tighter importance weighted lower bound, which is computed from multiple samples drawn from a reparameterizable Gumbel-Softmax sampler. A generic entropic mirror descent algorithm is applied to solve the resulting constrained variational inference task. The proposed method achieves the state-of-the-art performance on various popular scene graph generation benchmarks.

As a structured prediction task, scene graph generation aims to build a visually-grounded scene graph to explicitly model objects and their relationships in an input image. Currently, the mean field variational Bayesian framework is the de facto methodology used by the existing methods, in which the unconstrained inference step is often implemented by a message passing neural network. However, such formulation fails to explore other inference strategies, and largely ignores the more general constrained optimization models. In this paper, we present a constrained structure learning method, for which an explicit constrained variational inference objective is proposed. Instead of applying the ubiquitous message-passing strategy, a generic constrained optimization method - entropic mirror descent - is utilized to solve the constrained variational inference step. We validate the proposed generic model on various popular scene graph generation benchmarks and show that it outperforms the state-of-the-art methods.

The Symmetric Positive Definite (SPD) matrices have received wide attention for data representation in many scientific areas. Although there are many different attempts to develop effective deep architectures for data processing on the Riemannian manifold of SPD matrices, very few solutions explicitly mine the local geometrical information in deep SPD feature representations. Given the great success of local mechanisms in Euclidean methods, we argue that it is of utmost importance to ensure the preservation of local geometric information in the SPD networks. We first analyse the convolution operator commonly used for capturing local information in Euclidean deep networks from the perspective of a higher level of abstraction afforded by category theory. Based on this analysis, we define the local information in the SPD manifold and design a multi-scale submanifold block for mining local geometry. Experiments involving multiple visual tasks validate the effectiveness of our approach. The supplement and source code can be found in https://github.com/GitZH-Chen/MSNet.git.

We address the problem of multi-modal object tracking in video and explore various options of fusing the complementary information conveyed by the visible (RGB) and thermal infrared (TIR) modalities including pixel-level, feature-level and decision-level fusion. Specifically, different from the existing methods, paradigm of image fusion task is heeded for fusion at pixel level. Feature-level fusion is fulfilled by attention mechanism with channels excited optionally. Besides, at decision level, a novel fusion strategy is put forward since an effortless averaging configuration has shown the superiority. The effectiveness of the proposed decision-level fusion strategy owes to a number of innovative contributions, including a dynamic weighting of the RGB and TIR contributions and a linear template update operation. A variant of which produced the winning tracker at the Visual Object Tracking Challenge 2020 (VOT-RGBT2020). The concurrent exploration of innovative pixel- and feature-level fusion strategies highlights the advantages of the proposed decision-level fusion method. Extensive experimental results on three challenging datasets, \textit{i.e.}, GTOT, VOT-RGBT2019, and VOT-RGBT2020, demonstrate the effectiveness and robustness of the proposed method, compared to the state-of-the-art approaches. Code will be shared at \textcolor{blue}{\emph{https://github.com/Zhangyong-Tang/DFAT}., Comment: 13 pages, 10 figures

Scene graph generation aims to interpret an input image by explicitly modelling the potential objects and their relationships, which is predominantly solved by the message passing neural network models in previous methods. Currently, such approximation models generally assume the output variables are totally independent and thus ignore the informative structural higher-order interactions. This could lead to the inconsistent interpretations for an input image. In this paper, we propose a novel neural belief propagation method to generate the resulting scene graph. It employs a structural Bethe approximation rather than the mean field approximation to infer the associated marginals. To find a better bias-variance trade-off, the proposed model not only incorporates pairwise interactions but also higher order interactions into the associated scoring function. It achieves the state-of-the-art performance on various popular scene graph generation benchmarks.

Self-supervised pretraining is the method of choice for natural language processing models and is rapidly gaining popularity in many vision tasks. Recently, self-supervised pretraining has shown to outperform supervised pretraining for many downstream vision applications, marking a milestone in the area. This superiority is attributed to the negative impact of incomplete labelling of the training images, which convey multiple concepts, but are annotated using a single dominant class label. Although Self-Supervised Learning (SSL), in principle, is free of this limitation, the choice of pretext task facilitating SSL is perpetuating this shortcoming by driving the learning process towards a single concept output. This study aims to investigate the possibility of modelling all the concepts present in an image without using labels. In this aspect the proposed SSL frame-work MC-SSL0.0 is a step towards Multi-Concept Self-Supervised Learning (MC-SSL) that goes beyond modelling single dominant label in an image to effectively utilise the information from all the concepts present in it. MC-SSL0.0 consists of two core design concepts, group masked model learning and learning of pseudo-concept for data token using a momentum encoder (teacher-student) framework. The experimental results on multi-label and multi-class image classification downstream tasks demonstrate that MC-SSL0.0 not only surpasses existing SSL methods but also outperforms supervised transfer learning. The source code will be made publicly available for community to train on bigger corpus.

With the popularity of Transformer architectures in computer vision, the research focus has shifted towards developing computationally efficient designs. Window-based local attention is one of the major techniques being adopted in recent works. These methods begin with very small patch size and small embedding dimensions and then perform strided convolution (patch merging) in order to reduce the feature map size and increase embedding dimensions, hence, forming a pyramidal Convolutional Neural Network (CNN) like design. In this work, we investigate local and global information modelling in transformers by presenting a novel isotropic architecture that adopts local windows and special tokens, called Super tokens, for self-attention. Specifically, a single Super token is assigned to each image window which captures the rich local details for that window. These tokens are then employed for cross-window communication and global representation learning. Hence, most of the learning is independent of the image patches $(N)$ in the higher layers, and the class embedding is learned solely based on the Super tokens $(N/M^2)$ where $M^2$ is the window size. In standard image classification on Imagenet-1K, the proposed Super tokens based transformer (STT-S25) achieves 83.5\% accuracy which is equivalent to Swin transformer (Swin-B) with circa half the number of parameters (49M) and double the inference time throughput. The proposed Super token transformer offers a lightweight and promising backbone for visual recognition tasks.

A new multimodal biometric database designed and acquired within the framework of the European BioSecure Network of Excellence is presented. It is comprised of more than 600 individuals acquired simultaneously in three scenarios: 1) over the Internet, 2) in an office environment with desktop PC, and 3) in indoor/outdoor environments with mobile portable hardware. The three scenarios include a common part of audio/video data. Also, signature and fingerprint data have been acquired both with desktop PC and mobile portable hardware. Additionally, hand and iris data were acquired in the second scenario using desktop PC. Acquisition has been conducted by 11 European institutions. Additional features of the BioSecure Multimodal Database (BMDB) are: two acquisition sessions, several sensors in certain modalities, balanced gender and age distributions, multimodal realistic scenarios with simple and quick tasks per modality, cross-European diversity, availability of demographic data, and compatibility with other multimodal databases. The novel acquisition conditions of the BMDB allow us to perform new challenging research and evaluation of either monomodal or multimodal biometric systems, as in the recent BioSecure Multimodal Evaluation campaign. A description of this campaign including baseline results of individual modalities from the new database is also given. The database is expected to be available for research purposes through the BioSecure Association during 2008, Comment: Published at IEEE Transactions on Pattern Analysis and Machine Intelligence journal

Automatically verifying the identity of a person by means of biometrics is an important application in day-to-day activities such as accessing banking services and security control in airports. To increase the system reliability, several biometric devices are often used. Such a combined system is known as a multimodal biometric system. This paper reports a benchmarking study carried out within the framework of the BioSecure DS2 (Access Control) evaluation campaign organized by the University of Surrey, involving face, fingerprint, and iris biometrics for person authentication, targeting the application of physical access control in a medium-size establishment with some 500 persons. While multimodal biometrics is a well-investigated subject, there exists no benchmark for a fusion algorithm comparison. Working towards this goal, we designed two sets of experiments: quality-dependent and cost-sensitive evaluation. The quality-dependent evaluation aims at assessing how well fusion algorithms can perform under changing quality of raw images principally due to change of devices. The cost-sensitive evaluation, on the other hand, investigates how well a fusion algorithm can perform given restricted computation and in the presence of software and hardware failures, resulting in errors such as failure-to-acquire and failure-to-match. Since multiple capturing devices are available, a fusion algorithm should be able to handle this nonideal but nevertheless realistic scenario. In both evaluations, each fusion algorithm is provided with scores from each biometric comparison subsystem as well as the quality measures of both template and query data. The response to the call of the campaign proved very encouraging, with the submission of 22 fusion systems. To the best of our knowledge, this is the first attempt to benchmark quality-based multimodal fusion algorithms., Comment: Published at IEEE Transactions on Information Forensics and Security journal

Recent works have investigated deep learning models trained by optimising PAC-Bayes bounds, with priors that are learnt on subsets of the data. This combination has been shown to lead not only to accurate classifiers, but also to remarkably tight risk certificates, bearing promise towards self-certified learning (i.e. use all the data to learn a predictor and certify its quality). In this work, we empirically investigate the role of the prior. We experiment on 6 datasets with different strategies and amounts of data to learn data-dependent PAC-Bayes priors, and we compare them in terms of their effect on test performance of the learnt predictors and tightness of their risk certificate. We ask what is the optimal amount of data which should be allocated for building the prior and show that the optimum may be dataset dependent. We demonstrate that using a small percentage of the prior-building data for validation of the prior leads to promising results. We include a comparison of underparameterised and overparameterised models, along with an empirical study of different training objectives and regularisation strategies to learn the prior distribution.

In recent years, cross-media hashing technique has attracted increasing attention for its high computation efficiency and low storage cost. However, the existing approaches still have some limitations, which need to be explored. 1) A fixed hash length (e.g., 16bits or 32bits) is predefined before learning the binary codes. Therefore, these models need to be retrained when the hash length changes, that consumes additional computation power, reducing the scalability in practical applications. 2) Existing cross-modal approaches only explore the information in the original multimedia data to perform the hash learning, without exploiting the semantic information contained in the learned hash codes. To this end, we develop a novel Multiple hash cOdes jOint learNing method (MOON) for cross-media retrieval. Specifically, the developed MOON synchronously learns the hash codes with multiple lengths in a unified framework. Besides, to enhance the underlying discrimination, we combine the clues from the multimodal data, semantic labels and learned hash codes for hash learning. As far as we know, the proposed MOON is the first work to simultaneously learn different length hash codes without retraining in cross-media retrieval. Experiments on several databases show that our MOON can achieve promising performance, outperforming some recent competitive shallow and deep methods.

The Transformer architecture has witnessed a rapid development in recent years, outperforming the CNN architectures in many computer vision tasks, as exemplified by the Vision Transformers (ViT) for image classification. However, existing visual transformer models aim to extract semantic information for high-level tasks, such as classification and detection.These methods ignore the importance of the spatial resolution of the input image, thus sacrificing the local correlation information of neighboring pixels. In this paper, we propose a Patch Pyramid Transformer(PPT) to effectively address the above issues.Specifically, we first design a Patch Transformer to transform the image into a sequence of patches, where transformer encoding is performed for each patch to extract local representations. In addition, we construct a Pyramid Transformer to effectively extract the non-local information from the entire image. After obtaining a set of multi-scale, multi-dimensional, and multi-angle features of the original image, we design the image reconstruction network to ensure that the features can be reconstructed into the original input. To validate the effectiveness, we apply the proposed Patch Pyramid Transformer to image fusion tasks. The experimental results demonstrate its superior performance, compared to the state-of-the-art fusion approaches, achieving the best results on several evaluation indicators. Thanks to the underlying representational capacity of the PPT network, it can directly be applied to different image fusion tasks without redesigning or retraining the network., Comment: We have added more experiments and are improving the article, and will submit it to the journal. It will be updated

Self-supervised learning methods are gaining increasing traction in computer vision due to their recent success in reducing the gap with supervised learning. In natural language processing (NLP) self-supervised learning and transformers are already the methods of choice. The recent literature suggests that the transformers are becoming increasingly popular also in computer vision. So far, the vision transformers have been shown to work well when pretrained either using a large scale supervised data or with some kind of co-supervision, e.g. in terms of teacher network. These supervised pretrained vision transformers achieve very good results in downstream tasks with minimal changes. In this work we investigate the merits of self-supervised learning for pretraining image/vision transformers and then using them for downstream classification tasks. We propose Self-supervised vIsion Transformers (SiT) and discuss several self-supervised training mechanisms to obtain a pretext model. The architectural flexibility of SiT allows us to use it as an autoencoder and work with multiple self-supervised tasks seamlessly. We show that a pretrained SiT can be finetuned for a downstream classification task on small scale datasets, consisting of a few thousand images rather than several millions. The proposed approach is evaluated on standard datasets using common protocols. The results demonstrate the strength of the transformers and their suitability for self-supervised learning. We outperformed existing self-supervised learning methods by large margin. We also observed that SiT is good for few shot learning and also showed that it is learning useful representation by simply training a linear classifier on top of the learned features from SiT. Pretraining, finetuning, and evaluation codes will be available under: https://github.com/Sara-Ahmed/SiT.

In the image fusion field, the design of deep learning-based fusion methods is far from routine. It is invariably fusion-task specific and requires a careful consideration. The most difficult part of the design is to choose an appropriate strategy to generate the fused image for a specific task in hand. Thus, devising learnable fusion strategy is a very challenging problem in the community of image fusion. To address this problem, a novel end-to-end fusion network architecture (RFN-Nest) is developed for infrared and visible image fusion. We propose a residual fusion network (RFN) which is based on a residual architecture to replace the traditional fusion approach. A novel detail-preserving loss function, and a feature enhancing loss function are proposed to train RFN. The fusion model learning is accomplished by a novel two-stage training strategy. In the first stage, we train an auto-encoder based on an innovative nest connection (Nest) concept. Next, the RFN is trained using the proposed loss functions. The experimental results on public domain data sets show that, compared with the existing methods, our end-to-end fusion network delivers a better performance than the state-of-the-art methods in both subjective and objective evaluation. The code of our fusion method is available at https://github.com/hli1221/imagefusion-rfn-nest, Comment: Accepted by Information Fusion. 17 pages, 18 figures, 8 tables

Face recognition (FR) using deep convolutional neural networks (DCNNs) has seen remarkable success in recent years. One key ingredient of DCNN-based FR is the appropriate design of a loss function that ensures discrimination between various identities. The state-of-the-art (SOTA) solutions utilise normalised Softmax loss with additive and/or multiplicative margins. Despite being popular, these Softmax+margin based losses are not theoretically motivated and the effectiveness of a margin is justified only intuitively. In this work, we utilise an alternative framework that offers a more direct mechanism of achieving discrimination among the features of various identities. We propose a novel loss that is equivalent to a triplet loss with proxies and an implicit mechanism of hard-negative mining. We give theoretical justification that minimising the proposed loss ensures a minimum separability between all identities. The proposed loss is simple to implement and does not require heavy hyper-parameter tuning as in the SOTA solutions. We give empirical evidence that despite its simplicity, the proposed loss consistently achieves SOTA performance in various benchmarks for both high-resolution and low-resolution FR tasks.