Search

We propose and demonstrate an alternating Fourier and image domain filtering approach for feature extraction as an efficient alternative to build a vision backbone without using the computationally intensive attention. The performance among the lightweight models reaches the state-of-the-art level on ImageNet-1K classification, and improves downstream tasks on object detection and segmentation consistently as well. Our approach also serves as a new tool to compress vision transformers (ViTs).

While large multi-modal models (LMM) have shown notable progress in multi-modal tasks, their capabilities in tasks involving dense textual content remains to be fully explored. Dense text, which carries important information, is often found in documents, tables, and product descriptions. Understanding dense text enables us to obtain more accurate information, assisting in making better decisions. To further explore the capabilities of LMM in complex text tasks, we propose the DT-VQA dataset, with 170k question-answer pairs. In this paper, we conduct a comprehensive evaluation of GPT4V, Gemini, and various open-source LMMs on our dataset, revealing their strengths and weaknesses. Furthermore, we evaluate the effectiveness of two strategies for LMM: prompt engineering and downstream fine-tuning. We find that even with automatically labeled training datasets, significant improvements in model performance can be achieved. We hope that this research will promote the study of LMM in dense text tasks. Code will be released at https://github.com/Yuliang-Liu/MultimodalOCR.

Intercalation of epitaxial graphene on SiC(0001) with Sn results in a well-ordered 2D metallic Sn phase with a $(1\times1)$ structure at the interface between SiC substrate and quasi-freestanding graphene. The 2D\,Sn phase exhibits exotic electronic properties with Dirac-like and flat bands coexisting close to the Fermi level that exhibit both Zeeman- and Rashba-type spin splittings. Possible inter-layer interactions between the 2D\,Sn layer and graphene that may result in emerging electronic properties remain unexplored. We use time- and angle-resolved photoemission spectroscopy to reveal a surprisingly short-lived non-equilibrium carrier distribution inside the Dirac cone of graphene. Further, we find that the graphene $\pi$-band exhibits a transient down-shift that we attribute to charging of the graphene layer with holes. We interpret our results with support from density functional theory calculations of the graphene - 2D\,Sn heterostructure that reveal a substantial hybridization between graphene $\pi$-band and Sn $p_z$-states that opens up a $\sim230$\,meV band gap inside the Dirac cone and delocalizes the charge carriers over both the graphene and 2D\,Sn layers. Our results have important implications for the design of future ultrafast optoelectronic devices that may find applications in the fields of light harvesting and detection, as supercapacitors, or in novel quantum computing technologies., Comment: 19 pages, 5 figures

We present TextMonkey, a large multimodal model (LMM) tailored for text-centric tasks. Our approach introduces enhancement across several dimensions: By adopting Shifted Window Attention with zero-initialization, we achieve cross-window connectivity at higher input resolutions and stabilize early training; We hypothesize that images may contain redundant tokens, and by using similarity to filter out significant tokens, we can not only streamline the token length but also enhance the model's performance. Moreover, by expanding our model's capabilities to encompass text spotting and grounding, and incorporating positional information into responses, we enhance interpretability. It also learns to perform screenshot tasks through finetuning. Evaluation on 12 benchmarks shows notable improvements: 5.2% in Scene Text-Centric tasks (including STVQA, TextVQA, and OCRVQA), 6.9% in Document-Oriented tasks (such as DocVQA, InfoVQA, ChartVQA, DeepForm, Kleister Charity, and WikiTableQuestions), and 2.8% in Key Information Extraction tasks (comprising FUNSD, SROIE, and POIE). It outperforms in scene text spotting with a 10.9\% increase and sets a new standard on OCRBench, a comprehensive benchmark consisting of 29 OCR-related assessments, with a score of 561, surpassing previous open-sourced large multimodal models for document understanding. Code will be released at https://github.com/Yuliang-Liu/Monkey.

Scene text recognition (STR) is a challenging task that requires large-scale annotated data for training. However, collecting and labeling real text images is expensive and time-consuming, which limits the availability of real data. Therefore, most existing STR methods resort to synthetic data, which may introduce domain discrepancy and degrade the performance of STR models. To alleviate this problem, recent semi-supervised STR methods exploit unlabeled real data by enforcing character-level consistency regularization between weakly and strongly augmented views of the same image. However, these methods neglect word-level consistency, which is crucial for sequence recognition tasks. This paper proposes a novel semi-supervised learning method for STR that incorporates word-level consistency regularization from both visual and semantic aspects. Specifically, we devise a shortest path alignment module to align the sequential visual features of different views and minimize their distance. Moreover, we adopt a reinforcement learning framework to optimize the semantic similarity of the predicted strings in the embedding space. We conduct extensive experiments on several standard and challenging STR benchmarks and demonstrate the superiority of our proposed method over existing semi-supervised STR methods., Comment: Accepted by Pattern Recognition Letters

Scene text recognition is a rapidly developing field that faces numerous challenges due to the complexity and diversity of scene text, including complex backgrounds, diverse fonts, flexible arrangements, and accidental occlusions. In this paper, we propose a novel approach called Class-Aware Mask-guided feature refinement (CAM) to address these challenges. Our approach introduces canonical class-aware glyph masks generated from a standard font to effectively suppress background and text style noise, thereby enhancing feature discrimination. Additionally, we design a feature alignment and fusion module to incorporate the canonical mask guidance for further feature refinement for text recognition. By enhancing the alignment between the canonical mask feature and the text feature, the module ensures more effective fusion, ultimately leading to improved recognition performance. We first evaluate CAM on six standard text recognition benchmarks to demonstrate its effectiveness. Furthermore, CAM exhibits superiority over the state-of-the-art method by an average performance gain of 4.1% across six more challenging datasets, despite utilizing a smaller model size. Our study highlights the importance of incorporating canonical mask guidance and aligned feature refinement techniques for robust scene text recognition. The code is available at https://github.com/MelosY/CAM., Comment: Accepted by Pattern Recognition

Non-Hermitian optics has revealed a series of counterintuitive phenomena with profound implications for sensing, lasing, and light manipulation. While the non-Hermiticity of Hamitonians is well-recognized, recent advancements in non-Hermitian physics have broadened to include scattering matrices, uncovering phenomena such as simultaneous lasing and coherent perfect absorption (CPA), reflectionless scattering modes (RSMs), and coherent chaos control. Despite these developments, the investigation has predominantly focused on static and symmetric configurations, leaving the dynamic properties of non-Hermitian scattering in detuned systems largely unexplored. Bridging this gap, we extend certain stationary non-Hermitian scattering phenomena to detuned systems. We delve into the interplay between bi-directional RSMs and RSM exceptional points (EPs), and elucidate the global existence conditions for RSMs under detuning. Moreover, we introduces a novel category of EPs, characterized by the coalescence of transmission peaks, emerging independent with the presence of Hamiltonian EPs. The transmission EPs (TEPs) exhibit flat-top lineshape and can be extended to a square-shaped spectrum when detuning is involved, accompanied by a distinctive phase transition. Significantly, we demonstrate the applications of the TEPs in a one-dimensional coupled cavity system, engineered to enhance sensing robustness against environmental instabilities such as laser frequency drifts, which can exceed 10 MHz. This capability represents a substantial improvement over traditional sensing methods and an important improvement of fragile EP sensors. Our findings not only contribute to the broader understanding of non-Hermitian scattering phenomena but also paves the way for future advancements in non-Hermitian sensing technologies.

Large Multimodal Models (LMMs) have shown promise in vision-language tasks but struggle with high-resolution input and detailed scene understanding. Addressing these challenges, we introduce Monkey to enhance LMM capabilities. Firstly, Monkey processes input images by dividing them into uniform patches, each matching the size (e.g., 448x448) used in the original training of the well-trained vision encoder. Equipped with individual adapter for each patch, Monkey can handle higher resolutions up to 1344x896 pixels, enabling the detailed capture of complex visual information. Secondly, it employs a multi-level description generation method, enriching the context for scene-object associations. This two-part strategy ensures more effective learning from generated data: the higher resolution allows for a more detailed capture of visuals, which in turn enhances the effectiveness of comprehensive descriptions. Extensive ablative results validate the effectiveness of our designs. Additionally, experiments on 18 datasets further demonstrate that Monkey surpasses existing LMMs in many tasks like Image Captioning and various Visual Question Answering formats. Specially, in qualitative tests focused on dense text question answering, Monkey has exhibited encouraging results compared with GPT4V. Code is available at https://github.com/Yuliang-Liu/Monkey., Comment: CVPR 2024 Highlight

The experimental evidence for the $\alpha$ Boson gas state in the $^{11}$C+$^{12}$C$\rightarrow$$^{23}$Mg$^{\ast}$ fusion evaporation reaction is presented. By measuring the $\alpha$ emission spectrum with multiplicity 2 and 3, we provide insight into the existence of a three-body force among $\alpha$ particles. The observed spectrum exhibited distinct tails corresponding to $\alpha$ particles emitted in pairs and triplets consistent well with the model-calculations of AV18-UX and chiral effective field theory of NV2-3-la*, indicating the formation of $\alpha$ clusters with three-body force in the Boson gas state., Comment: 7 pages, 6 figures

The correlation of $\alpha\alpha$ was probed via measuring the transverse momentum $p_{T}$ and width $\delta p_{T}$ of one $\alpha$, for the first time, which represents the spatial and dynamical essentialities of the initial coupling state in $^{8}$Be nucleus. The weighted interaction vertex of 3$\alpha$ reflected by the magnitudes of their relative momentums and relative emission angles proves the isosceles triangle configuration for 3$\alpha$ at the high excited energy analogous Hoyle states., Comment: 8 pages, 9 figures

The enhancement of $J^{\pi}(T)$=3$^{+}$(0) state with isospin $T=0$ excited by the tensor force in the free $^{6}$Li nucleus has been observed, for the first time, relative to a shrinkable excitation in the $^{6}$Li cluster component inside its host nucleus. Comparatively, the excitation of $J^{\pi}(T)$=0$^{+}$(1) state with isospin $T=1$ for these two $^{6}$Li formations take on an approximately equal excitation strength. The mechanism of such tensor force effect was proposed due to the intensive nuclear medium role on isospin $T$=0 state., Comment: 6 pages, 4 figures

With their non-Abelian topological charges, real multi-bandgap systems challenge the conventional topological phase classifications. As the minimal sector of multi-bandgap systems, real triple degeneracies (RTPs), which serve as real 'Weyl points', lay the foundation for the research on real topological phases. However, experimental demonstration of physical systems with global band configurations consisting of multiple RTPs in crystals has not been reported. Here we present experimental evidence of RTPs in photonic meta-crystals, characterizing them using the Euler number, and establishing their connection with both Abelian and non-Abelian charges. By considering RTPs as the basic elements, we further propose the concept of a topological compound, akin to a chemical compound, where we find that certain phases are not topologically allowed. The topological classification of RTPs in crystals demonstrated in our work plays a similar role as the 'no-go' theorem in Weyl systems., Comment: 14 pages, 4 figures

Text recognition in the wild is a long-standing problem in computer vision. Driven by end-to-end deep learning, recent studies suggest vision and language processing are effective for scene text recognition. Yet, solving edit errors such as add, delete, or replace is still the main challenge for existing approaches. In fact, the content of the text and its audio are naturally corresponding to each other, i.e., a single character error may result in a clear different pronunciation. In this paper, we propose the AudioOCR, a simple yet effective probabilistic audio decoder for mel spectrogram sequence prediction to guide the scene text recognition, which only participates in the training phase and brings no extra cost during the inference stage. The underlying principle of AudioOCR can be easily applied to the existing approaches. Experiments using 7 previous scene text recognition methods on 12 existing regular, irregular, and occluded benchmarks demonstrate our proposed method can bring consistent improvement. More importantly, through our experimentation, we show that AudioOCR possesses a generalizability that extends to more challenging scenarios, including recognizing non-English text, out-of-vocabulary words, and text with various accents. Code will be available at https://github.com/wenwenyu/AudioOCR.

There has been a recent surge of interest in using light and sound as platforms for studying non-Abelian physics. Through a kaleidoscope of physical effects, light and sound provide diverse ways to manipulate their degrees of freedom to constitute the Hilbert space for demonstrating non-Abelian phenomena. The review aims to provide a timely and comprehensive account of this emerging topic. Starting from the foundation of matrix-valued geometric phases, we cover non-Abelian topological charges, non-Abelian gauge fields, non-Abelian braiding, non-Hermitian non-Abelian phenomena, and their realizations with photonics and acoustics. This topic is fast evolving at the intersection of atomic, molecular, optical physics, condensed matter physics, and mathematical physics, with fascinating prospects ahead.

Large models have recently played a dominant role in natural language processing and multimodal vision-language learning. However, their effectiveness in text-related visual tasks remains relatively unexplored. In this paper, we conducted a comprehensive evaluation of Large Multimodal Models, such as GPT4V and Gemini, in various text-related visual tasks including Text Recognition, Scene Text-Centric Visual Question Answering (VQA), Document-Oriented VQA, Key Information Extraction (KIE), and Handwritten Mathematical Expression Recognition (HMER). To facilitate the assessment of Optical Character Recognition (OCR) capabilities in Large Multimodal Models, we propose OCRBench, a comprehensive evaluation benchmark. OCRBench contains 29 datasets, making it the most comprehensive OCR evaluation benchmark available. Furthermore, our study reveals both the strengths and weaknesses of these models, particularly in handling multilingual text, handwritten text, non-semantic text, and mathematical expression recognition. Most importantly, the baseline results presented in this study could provide a foundational framework for the conception and assessment of innovative strategies targeted at enhancing zero-shot multimodal techniques. The evaluation pipeline and benchmark are available at https://github.com/Yuliang-Liu/MultimodalOCR.

In this paper, we propose a feature affinity (FA) assisted knowledge distillation (KD) method to improve quantization-aware training of deep neural networks (DNN). The FA loss on intermediate feature maps of DNNs plays the role of teaching middle steps of a solution to a student instead of only giving final answers in the conventional KD where the loss acts on the network logits at the output level. Combining logit loss and FA loss, we found that the quantized student network receives stronger supervision than from the labeled ground-truth data. The resulting FAQD is capable of compressing model on label-free data, which brings immediate practical benefits as pre-trained teacher models are readily available and unlabeled data are abundant. In contrast, data labeling is often laborious and expensive. Finally, we propose a fast feature affinity (FFA) loss that accurately approximates FA loss with a lower order of computational complexity, which helps speed up training for high resolution image input.

Nodal links are special configurations of band degeneracies in the momentum space, where nodal line branches encircle each other. In PT symmetric systems, nodal lines can be topologically characterized using the eigenvector frame rotations along an encircling loop and the linking structure can be described with non-Abelian frame charges interacting among adjacent bands. In this paper, we present a photonic multiple nodal links system, where non-adjacent band topology is proposed to characterize the hidden relation between nodal lines from non-adjacent band pairs. Through an orthogonal nodal chain, the nodal line from the lower two bands predicts the existence of nodal lines formed between the higher bands. We designed and fabricated a metamaterial, with which the multiple nodal links and non-adjacent band topology are experimentally demonstrated.

We propose the concept of 2D non-Abelian topological insulator which can explain the energy distributions of the edge states and corner states in systems with parity-time symmetry. From the viewpoint of non-Abelian band topology, we establish the constraints on the 2D Zak phase and polarization. We demonstrate that the corner states in some 2D systems can be explained as the boundary mode of the 1D edge states arising from the multi-band non-Abelian topology of the system. We also propose the use of off-diagonal Berry phase as complementary information to assist the prediction of edge states in non-Abelian topological insulators. Our work provides an alternative approach to study edge and corner modes and this idea can be extended to 3D systems.

It has been shown by many researchers that transformers perform as well as convolutional neural networks in many computer vision tasks. Meanwhile, the large computational costs of its attention module hinder further studies and applications on edge devices. Some pruning methods have been developed to construct efficient vision transformers, but most of them have considered image classification tasks only. Inspired by these results, we propose SiDT, a method for pruning vision transformer backbones on more complicated vision tasks like object detection, based on the search of transformer dimensions. Experiments on CIFAR-100 and COCO datasets show that the backbones with 20\% or 40\% dimensions/parameters pruned can have similar or even better performance than the unpruned models. Moreover, we have also provided the complexity analysis and comparisons with the previous pruning methods.