Scene text recognition (STR) methods have struggled to attain high accuracy and fast inference speed. Autoregressive (AR)-based models implement the recognition in a character-by-character manner, showing superiority in accuracy but with slow inference speed. Alternatively, parallel decoding (PD)-based models infer all characters in a single decoding pass, offering faster inference speed but generally worse accuracy. We first present an empirical study of AR decoding in STR, and discover that the AR decoder not only models linguistic context, but also provides guidance on visual context perception. Consequently, we propose Context Perception Parallel Decoder (CPPD) to predict the character sequence in a PD pass. CPPD devises a character counting module to infer the occurrence count of each character, and a character ordering module to deduce the content-free reading order and placeholders. Meanwhile, the character prediction task associates the placeholders with characters. They together build a comprehensive recognition context. We construct a series of CPPD models and also plug the proposed modules into existing STR decoders. Experiments on both English and Chinese benchmarks demonstrate that the CPPD models achieve highly competitive accuracy while running approximately 8x faster than their AR-based counterparts. Moreover, the plugged models achieve significant accuracy improvements. Code is at \href{https://github.com/PaddlePaddle/PaddleOCR/blob/dygraph/doc/doc_en/algorithm_rec_cppd_en.md}{this https URL}.
The efficacy of building footprint segmentation from remotely sensed images has been hindered by model transfer effectiveness. Many existing building segmentation methods were developed upon the encoder-decoder architecture of U-Net, in which the encoder is finetuned from the newly developed backbone networks that are pre-trained on ImageNet. However, the heavy computational burden of the existing decoder designs hampers the successful transfer of these modern encoder networks to remote sensing tasks. Even the widely-adopted deep supervision strategy fails to mitigate these challenges due to its invalid loss in hybrid regions where foreground and background pixels are intermixed. In this paper, we conduct a comprehensive evaluation of existing decoder network designs for building footprint segmentation and propose an efficient framework denoted as BFSeg to enhance learning efficiency and effectiveness. Specifically, a densely-connected coarse-to-fine feature fusion decoder network that facilitates easy and fast feature fusion across scales is proposed. Moreover, considering the invalidity of hybrid regions in the down-sampled ground truth during the deep supervision process, we present a lenient deep supervision and distillation strategy that enables the network to learn proper knowledge from deep supervision. Building upon these advancements, we have developed a new family of building segmentation networks, which consistently surpass prior works with outstanding performance and efficiency across a wide range of newly developed encoder networks.
Parameter Efficient Tuning (PET) has gained attention for reducing the number of parameters while maintaining performance and providing better hardware resource savings, but few studies investigate dense prediction tasks and interaction between modalities. In this paper, we do an investigation of efficient tuning problems on referring image segmentation. We propose a novel adapter called Bridger to facilitate cross-modal information exchange and inject task-specific information into the pre-trained model. We also design a lightweight decoder for image segmentation. Our approach achieves comparable or superior performance with only 1.61\% to 3.38\% backbone parameter updates, evaluated on challenging benchmarks. The code is available at \url{https://github.com/kkakkkka/ETRIS}.
Fundus photography is an essential examination for clinical and differential diagnosis of fundus diseases. Recently, Ultra-Wide-angle Fundus (UWF) techniques, UWF Fluorescein Angiography (UWF-FA) and UWF Scanning Laser Ophthalmoscopy (UWF-SLO) have been gradually put into use. However, Fluorescein Angiography (FA) and UWF-FA require injecting sodium fluorescein which may have detrimental influences. To avoid negative impacts, cross-modality medical image generation algorithms have been proposed. Nevertheless, current methods in fundus imaging could not produce high-resolution images and are unable to capture tiny vascular lesion areas. This paper proposes a novel conditional generative adversarial network (UWAT-GAN) to synthesize UWF-FA from UWF-SLO. Using multi-scale generators and a fusion module patch to better extract global and local information, our model can generate high-resolution images. Moreover, an attention transmit module is proposed to help the decoder learn effectively. Besides, a supervised approach is used to train the network using multiple new weighted losses on different scales of data. Experiments on an in-house UWF image dataset demonstrate the superiority of the UWAT-GAN over the state-of-the-art methods. The source code is available at: https://github.com/Tinysqua/UWAT-GAN.
Source code summarization is the task of writing natural language descriptions of source code behavior. Code summarization underpins software documentation for programmers. Short descriptions of code help programmers understand the program quickly without having to read the code itself. Lately, neural source code summarization has emerged as the frontier of research into automated code summarization techniques. By far the most popular targets for summarization are program subroutines. The idea, in a nutshell, is to train an encoder-decoder neural architecture using large sets of examples of subroutines extracted from code repositories. The encoder represents the code and the decoder represents the summary. However, most current approaches attempt to treat the subroutine as a single unit. For example, by taking the entire subroutine as input to a Transformer or RNN-based encoder. But code behavior tends to depend on the flow from statement to statement. Normally dynamic analysis may shed light on this flow, but dynamic analysis on hundreds of thousands of examples in large datasets is not practical. In this paper, we present a statement-based memory encoder that learns the important elements of flow during training, leading to a statement-based subroutine representation without the need for dynamic analysis. We implement our encoder for code summarization and demonstrate a significant improvement over the state-of-the-art.
Weakly-supervised change detection (WSCD) aims to detect pixel-level changes with only image-level annotations. Owing to its label efficiency, WSCD is drawing increasing attention recently. However, current WSCD methods often encounter the challenge of change missing and fabricating, i.e., the inconsistency between image-level annotations and pixel-level predictions. Specifically, change missing refer to the situation that the WSCD model fails to predict any changed pixels, even though the image-level label indicates changed, and vice versa for change fabricating. To address this challenge, in this work, we leverage global-scale and local-scale priors in WSCD and propose two components: a Dilated Prior (DP) decoder and a Label Gated (LG) constraint. The DP decoder decodes samples with the changed image-level label, skips samples with the unchanged label, and replaces them with an all-unchanged pixel-level label. The LG constraint is derived from the correspondence between changed representations and image-level labels, penalizing the model when it mispredicts the change status. Additionally, we develop TransWCD, a simple yet powerful transformer-based model, showcasing the potential of weakly-supervised learning in change detection. By integrating the DP decoder and LG constraint into TransWCD, we form TransWCD-DL. Our proposed TransWCD and TransWCD-DL achieve significant +6.33% and +9.55% F1 score improvements over the state-of-the-art methods on the WHU-CD dataset, respectively. Some performance metrics even exceed several fully-supervised change detection (FSCD) competitors. Code will be available at https://github.com/zhenghuizhao/TransWCD.
Object localization in general environments is a fundamental part of vision systems. While dominating on the COCO benchmark, recent Transformer-based detection methods are not competitive in diverse domains. Moreover, these methods still struggle to very accurately estimate the object bounding boxes in complex environments. We introduce Cascade-DETR for high-quality universal object detection. We jointly tackle the generalization to diverse domains and localization accuracy by proposing the Cascade Attention layer, which explicitly integrates object-centric information into the detection decoder by limiting the attention to the previous box prediction. To further enhance accuracy, we also revisit the scoring of queries. Instead of relying on classification scores, we predict the expected IoU of the query, leading to substantially more well-calibrated confidences. Lastly, we introduce a universal object detection benchmark, UDB10, that contains 10 datasets from diverse domains. While also advancing the state-of-the-art on COCO, Cascade-DETR substantially improves DETR-based detectors on all datasets in UDB10, even by over 10 mAP in some cases. The improvements under stringent quality requirements are even more pronounced. Our code and models will be released at https://github.com/SysCV/cascade-detr.
We view variational autoencoders (VAE) as decoder-encoder pairs, which map distributions in the data space to distributions in the latent space and vice versa. The standard learning approach for VAEs is the maximisation of the evidence lower bound (ELBO). It is asymmetric in that it aims at learning a latent variable model while using the encoder as an auxiliary means only. Moreover, it requires a closed form a-priori latent distribution. This limits its applicability in more complex scenarios, such as general semi-supervised learning and employing complex generative models as priors. We propose a Nash equilibrium learning approach, which is symmetric with respect to the encoder and decoder and allows learning VAEs in situations where both the data and the latent distributions are accessible only by sampling. The flexibility and simplicity of this approach allows its application to a wide range of learning scenarios and downstream tasks.
Most end-to-end (E2E) speech recognition models are composed of encoder and decoder blocks that perform acoustic and language modeling functions. Pretrained large language models (LLMs) have the potential to improve the performance of E2E ASR. However, integrating a pretrained language model into an E2E speech recognition model has shown limited benefits due to the mismatches between text-based LLMs and those used in E2E ASR. In this paper, we explore an alternative approach by adapting a pretrained LLMs to speech. Our experiments on fully-formatted E2E ASR transcription tasks across various domains demonstrate that our approach can effectively leverage the strengths of pretrained LLMs to produce more readable ASR transcriptions. Our model, which is based on the pretrained large language models with either an encoder-decoder or decoder-only structure, surpasses strong ASR models such as Whisper, in terms of recognition error rate, considering formats like punctuation and capitalization as well.
Optical information processing and computing can potentially offer enhanced performance, scalability and energy efficiency. However, achieving nonlinearity-a critical component of computation-remains challenging in the optical domain. Here we introduce a design that leverages a multiple-scattering cavity to passively induce optical nonlinear random mapping with a continuous-wave laser at a low power. Each scattering event effectively mixes information from different areas of a spatial light modulator, resulting in a highly nonlinear mapping between the input data and output pattern. We demonstrate that our design retains vital information even when the readout dimensionality is reduced, thereby enabling optical data compression. This capability allows our optical platforms to offer efficient optical information processing solutions across applications. We demonstrate our design's efficacy across tasks, including classification, image reconstruction, keypoint detection and object detection, all of which are achieved through optical data compression combined with a digital decoder. In particular, high performance at extreme compression ratios is observed in real-time pedestrian detection. Our findings open pathways for novel algorithms and unconventional architectural designs for optical computing.
This paper shows that time series forecasting Transformer (TSFT) suffers from severe over-fitting problem caused by improper initialization method of unknown decoder inputs, esp. when handling non-stationary time series. Based on this observation, we propose GBT, a novel two-stage Transformer framework with Good Beginning. It decouples the prediction process of TSFT into two stages, including Auto-Regression stage and Self-Regression stage to tackle the problem of different statistical properties between input and prediction sequences.Prediction results of Auto-Regression stage serve as a Good Beginning, i.e., a better initialization for inputs of Self-Regression stage. We also propose Error Score Modification module to further enhance the forecasting capability of the Self-Regression stage in GBT. Extensive experiments on seven benchmark datasets demonstrate that GBT outperforms SOTA TSFTs (FEDformer, Pyraformer, ETSformer, etc.) and many other forecasting models (SCINet, N-HiTS, etc.) with only canonical attention and convolution while owning less time and space complexity. It is also general enough to couple with these models to strengthen their forecasting capability. The source code is available at: https://github.com/OrigamiSL/GBT
Few-shot video segmentation is the task of delineating a specific novel class in a query video using few labelled support images. Typical approaches compare support and query features while limiting comparisons to a single feature layer and thereby ignore potentially valuable information. We present a meta-learned Multiscale Memory Comparator (MMC) for few-shot video segmentation that combines information across scales within a transformer decoder. Typical multiscale transformer decoders for segmentation tasks learn a compressed representation, their queries, through information exchange across scales. Unlike previous work, we instead preserve the detailed feature maps during across scale information exchange via a multiscale memory transformer decoding to reduce confusion between the background and novel class. Integral to the approach, we investigate multiple forms of information exchange across scales in different tasks and provide insights with empirical evidence on which to use in each task. The overall comparisons among query and support features benefit from both rich semantics and precise localization. We demonstrate our approach primarily on few-shot video object segmentation and an adapted version on the fully supervised counterpart. In all cases, our approach outperforms the baseline and yields state-of-the-art performance. Our code is publicly available at https://github.com/MSiam/MMC-MultiscaleMemory.
Grounded Situation Recognition (GSR) is capable of recognizing and interpreting visual scenes in a contextually intuitive way, yielding salient activities (verbs) and the involved entities (roles) depicted in images. In this work, we focus on the application of GSR in assisting people with visual impairments (PVI). However, precise localization information of detected objects is often required to navigate their surroundings confidently and make informed decisions. For the first time, we propose an Open Scene Understanding (OpenSU) system that aims to generate pixel-wise dense segmentation masks of involved entities instead of bounding boxes. Specifically, we build our OpenSU system on top of GSR by additionally adopting an efficient Segment Anything Model (SAM). Furthermore, to enhance the feature extraction and interaction between the encoder-decoder structure, we construct our OpenSU system using a solid pure transformer backbone to improve the performance of GSR. In order to accelerate the convergence, we replace all the activation functions within the GSR decoders with GELU, thereby reducing the training duration. In quantitative analysis, our model achieves state-of-the-art performance on the SWiG dataset. Moreover, through field testing on dedicated assistive technology datasets and application demonstrations, the proposed OpenSU system can be used to enhance scene understanding and facilitate the independent mobility of people with visual impairments. Our code will be available at https://github.com/RuipingL/OpenSU.
Skeleton sequence representation learning has shown great advantages for action recognition due to its promising ability to model human joints and topology. However, the current methods usually require sufficient labeled data for training computationally expensive models, which is labor-intensive and time-consuming. Moreover, these methods ignore how to utilize the fine-grained dependencies among different skeleton joints to pre-train an efficient skeleton sequence learning model that can generalize well across different datasets. In this paper, we propose an efficient skeleton sequence learning framework, named Skeleton Sequence Learning (SSL). To comprehensively capture the human pose and obtain discriminative skeleton sequence representation, we build an asymmetric graph-based encoder-decoder pre-training architecture named SkeletonMAE, which embeds skeleton joint sequence into Graph Convolutional Network (GCN) and reconstructs the masked skeleton joints and edges based on the prior human topology knowledge. Then, the pre-trained SkeletonMAE encoder is integrated with the Spatial-Temporal Representation Learning (STRL) module to build the SSL framework. Extensive experimental results show that our SSL generalizes well across different datasets and outperforms the state-of-the-art self-supervised skeleton-based action recognition methods on FineGym, Diving48, NTU 60 and NTU 120 datasets. Additionally, we obtain comparable performance to some fully supervised methods. The code is avaliable at https://github.com/HongYan1123/SkeletonMAE.
Transformer and its variants have been widely used for medical image segmentation. However, the large number of parameter and computational load of these models make them unsuitable for mobile health applications. To address this issue, we propose a more efficient approach, the Efficient Group Enhanced UNet (EGE-UNet). We incorporate a Group multi-axis Hadamard Product Attention module (GHPA) and a Group Aggregation Bridge module (GAB) in a lightweight manner. The GHPA groups input features and performs Hadamard Product Attention mechanism (HPA) on different axes to extract pathological information from diverse perspectives. The GAB effectively fuses multi-scale information by grouping low-level features, high-level features, and a mask generated by the decoder at each stage. Comprehensive experiments on the ISIC2017 and ISIC2018 datasets demonstrate that EGE-UNet outperforms existing state-of-the-art methods. In short, compared to the TransFuse, our model achieves superior segmentation performance while reducing parameter and computation costs by 494x and 160x, respectively. Moreover, to our best knowledge, this is the first model with a parameter count limited to just 50KB. Our code is available at https://github.com/JCruan519/EGE-UNet.
Conditional spatial queries are recently introduced into DEtection TRansformer (DETR) to accelerate convergence. In DAB-DETR, such queries are modulated by the so-called conditional linear projection at each decoder stage, aiming to search for positions of interest such as the four extremities of the box. Each decoder stage progressively updates the box by predicting the anchor box offsets, while in cross-attention only the box center is informed as the reference point. The use of only box center, however, leaves the width and height of the previous box unknown to the current stage, which hinders accurate prediction of offsets. We argue that the explicit use of the entire box information in cross-attention matters. In this work, we propose Box Agent to condense the box into head-specific agent points. By replacing the box center with the agent point as the reference point in each head, the conditional cross-attention can search for positions from a more reasonable starting point by considering the full scope of the previous box, rather than always from the previous box center. This significantly reduces the burden of the conditional linear projection. Experimental results show that the box agent leads to not only faster convergence but also improved detection performance, e.g., our single-scale model achieves $44.2$ AP with ResNet-50 based on DAB-DETR. Our Box Agent requires minor modifications to the code and has negligible computational workload. Code is available at https://github.com/tiny-smart/box-detr.
We introduce the notion of point affiliation into feature upsampling. By abstracting a feature map into non-overlapped semantic clusters formed by points of identical semantic meaning, feature upsampling can be viewed as point affiliation -- designating a semantic cluster for each upsampled point. In the framework of kernel-based dynamic upsampling, we show that an upsampled point can resort to its low-res decoder neighbors and high-res encoder point to reason the affiliation, conditioned on the mutual similarity between them. We therefore present a generic formulation for generating similarity-aware upsampling kernels and prove that such kernels encourage not only semantic smoothness but also boundary sharpness. This formulation constitutes a novel, lightweight, and universal upsampling solution, Similarity-Aware Point Affiliation (SAPA). We show its working mechanism via our preliminary designs with window-shape kernel. After probing the limitations of the designs on object detection, we reveal additional insights for upsampling, leading to SAPA with the dynamic kernel shape. Extensive experiments demonstrate that SAPA outperforms prior upsamplers and invites consistent performance improvements on a number of dense prediction tasks, including semantic segmentation, object detection, instance segmentation, panoptic segmentation, image matting, and depth estimation. Code is made available at: https://github.com/tiny-smart/sapa
Diffusion models have recently emerged as powerful generative priors for solving inverse problems. However, training diffusion models in the pixel space are both data-intensive and computationally demanding, which restricts their applicability as priors for high-dimensional real-world data such as medical images. Latent diffusion models, which operate in a much lower-dimensional space, offer a solution to these challenges. However, incorporating latent diffusion models to solve inverse problems remains a challenging problem due to the nonlinearity of the encoder and decoder. To address these issues, we propose \textit{ReSample}, an algorithm that can solve general inverse problems with pre-trained latent diffusion models. Our algorithm incorporates data consistency by solving an optimization problem during the reverse sampling process, a concept that we term as hard data consistency. Upon solving this optimization problem, we propose a novel resampling scheme to map the measurement-consistent sample back onto the noisy data manifold and theoretically demonstrate its benefits. Lastly, we apply our algorithm to solve a wide range of linear and nonlinear inverse problems in both natural and medical images, demonstrating that our approach outperforms existing state-of-the-art approaches, including those based on pixel-space diffusion models.
Simulation forms the backbone of modern self-driving development. Simulators help develop, test, and improve driving systems without putting humans, vehicles, or their environment at risk. However, simulators face a major challenge: They rely on realistic, scalable, yet interesting content. While recent advances in rendering and scene reconstruction make great strides in creating static scene assets, modeling their layout, dynamics, and behaviors remains challenging. In this work, we turn to language as a source of supervision for dynamic traffic scene generation. Our model, LCTGen, combines a large language model with a transformer-based decoder architecture that selects likely map locations from a dataset of maps, and produces an initial traffic distribution, as well as the dynamics of each vehicle. LCTGen outperforms prior work in both unconditional and conditional traffic scene generation in terms of realism and fidelity. Code and video will be available at https://ariostgx.github.io/lctgen.
Automatic and accurate segmentation of colon polyps is essential for early diagnosis of colorectal cancer. Advanced deep learning models have shown promising results in polyp segmentation. However, they still have limitations in representing multi-scale features and generalization capability. To address these issues, this paper introduces RaBiT, an encoder-decoder model that incorporates a lightweight Transformer-based architecture in the encoder to model multiple-level global semantic relationships. The decoder consists of several bidirectional feature pyramid layers with reverse attention modules to better fuse feature maps at various levels and incrementally refine polyp boundaries. We also propose ideas to lighten the reverse attention module and make it more suitable for multi-class segmentation. Extensive experiments on several benchmark datasets show that our method outperforms existing methods across all datasets while maintaining low computational complexity. Moreover, our method demonstrates high generalization capability in cross-dataset experiments, even when the training and test sets have different characteristics.