Semantic Regularization Research Articles

Investigation of Visual Language Landscape of Tourist Attractions from Multimodal Perspective

With the development of economic globalization, the tourism industry has been welcomed by the public. The visual language landscape of tourist attractions can not only assist tourists to play and watch the project, but if it is properly planned, the language landscape can also become a major feature and highlight of the scenic spot. Therefore, how to set up and construct the visual language landscape of tourist attractions is a problem that needs to be considered in each region. In response to the above problems, on the basis of understanding the concept types of the visual language landscape of tourist attractions, this article conducts in-depth research and investigation on the visual language landscape of tourist attractions, combining the evaluation dataset in the multimodal perspective and the Convolutional Neural Network (CNN) –Recurrent Neural Network (RNN) model based on semantic regularization. This article conducted a comparative experiment on each model on the NUS-WIDE dataset and the MS-COCO dataset. The experimental results showed that it was crucial to give full play to the expressive power of the CNN. Compared to the NUS-WIDE dataset, the MS-COCO dataset brought less additional boost by leveraging social media tags. The CIDEr score of the CNN-RNN model based on semantic regularization was improved by 11.4%, which placed the foundation for the investigation and analysis of the linguistic landscape of tourist attractions.

Semantic regularization of electromagnetic inverse problems

Solving ill-posed inverse problems typically requires regularization based on prior knowledge. To date, only prior knowledge that is formulated mathematically (e.g., sparsity of the unknown) or implicitly learned from quantitative data can be used for regularization. Thereby, semantically formulated prior knowledge derived from human reasoning and recognition is excluded. Here, we introduce and demonstrate the concept of semantic regularization based on a pre-trained large language model to overcome this vexing limitation. We study the approach, first, numerically in a prototypical 2D inverse scattering problem, and, second, experimentally in 3D and 4D compressive microwave imaging problems based on programmable metasurfaces. We highlight that semantic regularization enables new forms of highly-sought privacy protection for applications like smart homes, touchless human-machine interaction and security screening: selected subjects in the scene can be concealed, or their actions and postures can be altered in the reconstruction by manipulating the semantic prior with suitable language-based control commands.

Dynamic heterogeneous federated learning with multi-level prototypes

Federated learning shows promise as a privacy-preserving collaborative learning technique. Existing research mainly focuses on skewing the class distribution across clients. However, most approaches suffer from catastrophic forgetting and classifier shift, mainly when the global distribution of all classes is extremely unbalanced and the data distribution of the client dynamically evolves over time. In this paper, we study the Dynamic Heterogeneous Federated Learning, which addresses the practical scenario where heterogeneous data distributions exist among different clients and dynamic tasks within the client. Accordingly, we propose a novel federated learning framework named Federated Multi-Level Prototypes and design federated multi-level regularizations. To mitigate classifier shift, we construct semantic prototypes to provide fruitful generalization knowledge. To maintain the model stability and consistency convergence, three regularizations are introduced as training losses, i.e., prototype-based regularization, semantic prototype-based regularization, and federated inter-task regularization. Extensive experiments show that the proposed method achieves advanced performance in various settings.

Beyond Prototypes: Semantic Anchor Regularization for Better Representation Learning

One of the ultimate goals of representation learning is to achieve compactness within a class and well-separability between classes. Many outstanding metric-based and prototype-based methods following the Expectation-Maximization paradigm, have been proposed for this objective. However, they inevitably introduce biases into the learning process, particularly with long-tail distributed training data. In this paper, we reveal that the class prototype is not necessarily to be derived from training features and propose a novel perspective to use pre-defined class anchors serving as feature centroid to unidirectionally guide feature learning. However, the pre-defined anchors may have a large semantic distance from the pixel features, which prevents them from being directly applied. To address this issue and generate feature centroid independent from feature learning, a simple yet effective Semantic Anchor Regularization (SAR) is proposed. SAR ensures the inter-class separability of semantic anchors in the semantic space by employing a classifier-aware auxiliary cross-entropy loss during training via disentanglement learning. By pulling the learned features to these semantic anchors, several advantages can be attained: 1) the intra-class compactness and naturally inter-class separability, 2) induced bias or errors from feature learning can be avoided, and 3) robustness to the long-tailed problem. The proposed SAR can be used in a plug-and-play manner in the existing models. Extensive experiments demonstrate that the SAR performs better than previous sophisticated prototype-based methods. The implementation is available at https://github.com/geyanqi/SAR.

NeBLa: Neural Beer-Lambert for 3D Reconstruction of Oral Structures from Panoramic Radiographs

Panoramic radiography (Panoramic X-ray, PX) is a widely used imaging modality for dental examination. However, PX only provides a flattened 2D image, lacking in a 3D view of the oral structure. In this paper, we propose NeBLa (Neural Beer-Lambert) to estimate 3D oral structures from real-world PX. NeBLa tackles full 3D reconstruction for varying subjects (patients) where each reconstruction is based only on a single panoramic image. We create an intermediate representation called simulated PX (SimPX) from 3D Cone-beam computed tomography (CBCT) data based on the Beer-Lambert law of X-ray rendering and rotational principles of PX imaging. SimPX aims at not only truthfully simulating PX, but also facilitates the reverting process back to 3D data. We propose a novel neural model based on ray tracing which exploits both global and local input features to convert SimPX to 3D output. At inference, a real PX image is translated to a SimPX-style image with semantic regularization, and the translated image is processed by generation module to produce high-quality outputs. Experiments show that NeBLa outperforms prior state-of-the-art in reconstruction tasks both quantitatively and qualitatively. Unlike prior methods, NeBLa does not require any prior information such as the shape of dental arches, nor the matched PX-CBCT dataset for training, which is difficult to obtain in clinical practice. Our code is available at https://github.com/sihwa-park/nebla.

VLM2Scene: Self-Supervised Image-Text-LiDAR Learning with Foundation Models for Autonomous Driving Scene Understanding

Vision and language foundation models (VLMs) have showcased impressive capabilities in 2D scene understanding. However, their latent potential in elevating the understanding of 3D autonomous driving scenes remains untapped. In this paper, we propose VLM2Scene, which exploits the potential of VLMs to enhance 3D self-supervised representation learning through our proposed image-text-LiDAR contrastive learning strategy. Specifically, in the realm of autonomous driving scenes, the inherent sparsity of LiDAR point clouds poses a notable challenge for point-level contrastive learning methods. This method often grapples with limitations tied to a restricted receptive field and the presence of noisy points. To tackle this challenge, our approach emphasizes region-level learning, leveraging regional masks without semantics derived from the vision foundation model. This approach capitalizes on valuable contextual information to enhance the learning of point cloud representations. First, we introduce Region Caption Prompts to generate fine-grained language descriptions for the corresponding regions, utilizing the language foundation model. These region prompts then facilitate the establishment of positive and negative text-point pairs within the contrastive loss framework. Second, we propose a Region Semantic Concordance Regularization, which involves a semantic-filtered region learning and a region semantic assignment strategy. The former aims to filter the false negative samples based on the semantic distance, and the latter mitigates potential inaccuracies in pixel semantics, thereby enhancing overall semantic consistency. Extensive experiments on representative autonomous driving datasets demonstrate that our self-supervised method significantly outperforms other counterparts. Codes are available at https://github.com/gbliao/VLM2Scene.

Semantic-Aware Data Augmentation for Text-to-Image Synthesis

Data augmentation has been recently leveraged as an effective regularizer in various vision-language deep neural networks. However, in text-to-image synthesis (T2Isyn), current augmentation wisdom still suffers from the semantic mismatch between augmented paired data. Even worse, semantic collapse may occur when generated images are less semantically constrained. In this paper, we develop a novel Semantic-aware Data Augmentation (SADA) framework dedicated to T2Isyn. In particular, we propose to augment texts in the semantic space via an Implicit Textual Semantic Preserving Augmentation, in conjunction with a specifically designed Image Semantic Regularization Loss as Generated Image Semantic Conservation, to cope well with semantic mismatch and collapse. As one major contribution, we theoretically show that Implicit Textual Semantic Preserving Augmentation can certify better text-image consistency while Image Semantic Regularization Loss regularizing the semantics of generated images would avoid semantic collapse and enhance image quality. Extensive experiments validate that SADA enhances text-image consistency and improves image quality significantly in T2Isyn models across various backbones. Especially, incorporating SADA during the tuning process of Stable Diffusion models also yields performance improvements.

Zero-Shot Learning With Attentive Region Embedding and Enhanced Semantics.

The performance of zero-shot learning (ZSL) can be improved progressively by learning better features and generating pseudosamples for unseen classes. Existing ZSL works typically learn feature extractors and generators independently, which may shift the unseen samples away from their real distribution and suffers from the domain bias problem. In this article, to tackle this challenge, we propose a variational autoencoder (VAE)-based framework, that is, joint Attentive Region Embedding with Enhanced Semantics (AREES), which is tailored to advance the zero-shot recognition. Specifically, AREES is end-to-end trainable and consists of three network branches: 1) attentive region embedding is used to learn the semantic-guided visual features by the attention mechanism (AM); 2) a decomposition structure and a semantic pivot regularization are used to extract enhanced semantics; and 3) a multimodal VAE (mVAE) with the cross-reconstruction loss and the distribution alignment loss is used to obtain a shared latent embedding space of visual features and semantics. Finally, features' extraction and features' generation are optimized together in AREES to address the domain shift problem to a large extent. The comprehensive evaluations on six benchmarks, including the ImageNet, demonstrate the superiority of the proposed model over its state-of-the-art counterparts.

Self-supervised multi-scale semantic consistency regularization for unsupervised image-to-image translation

Unsupervised image-to-image translation aims to learn a domain mapping function that preserves the semantics of an input image while adapting its style to target domains without paired data. However, if there is a large semantic mismatch between the source and target domains, current methods often suffer from semantics distortion. Based on dense self-supervised representation learning, a novel Multi-Scale Semantic Consistency Regularization (MSSCR) is presented to alleviate the semantic distortion and enable the generator to produce images with realistic local semantics and consistent structures. Both local and global multi-scale representations are learned by the MSSCR during training the different layers of a discriminator. Concretely, MSSCR operates by sliding a fixed-size window over the overlapping region between a pair of views cropped from a single real image, aligning these areas with their corresponding multi-scale representation regions extracted from the discriminator, and then maximizing the similarity of representations between positive pairs. Qualitative and quantitative experiments demonstrate the superiority of MSSCR on image-to-image translation and image generation tasks.

SAB Net: A Semantic Attention Boosting Framework for Semantic Segmentation.

Semantic segmentation has achieved great progress by effectively fusing features of contextual information. In this article, we propose an end-to-end semantic attention boosting (SAB) framework to adaptively fuse the contextual information iteratively across layers with semantic regularization. Specifically, we first propose a pixelwise semantic attention (SAP) block, with a semantic metric representing the pixelwise category relationship, to aggregate the nonlocal contextual information. In addition, we improve the computation complexity of SAP block from O(n²) to O(n) for images with size n. Second, we present a categorywise semantic attention (SAC) block to adaptively balance the nonlocal contextual dependencies and the local consistency with a categorywise weight, overcoming the contextual information confusion caused by the feature imbalance within intra-category. Furthermore, we propose the SAB module to refine the segmentation with SAC and SAP blocks. By applying the SAB module iteratively across layers, our model shrinks the semantic gap and enhances the structure reasoning by fully utilizing the coarse segmentation information. Extensive quantitative evaluations demonstrate that our method significantly improves the segmentation results and achieves superior performance on the PASCAL VOC 2012, Cityscapes, PASCAL Context, and ADE20K datasets.

Universal Prototype Transport for Zero-Shot Action Recognition and Localization

This work addresses the problem of recognizing action categories in videos when no training examples are available. The current state-of-the-art enables such a zero-shot recognition by learning universal mappings from videos to a semantic space, either trained on large-scale seen actions or on objects. While effective, we find that universal action and object mappings are biased to specific regions in the semantic space. These biases lead to a fundamental problem: many unseen action categories are simply never inferred during testing. For example on UCF-101, a quarter of the unseen actions are out of reach with a state-of-the-art universal action model. To that end, this paper introduces universal prototype transport for zero-shot action recognition. The main idea is to re-position the semantic prototypes of unseen actions by matching them to the distribution of all test videos. For universal action models, we propose to match distributions through a hyperspherical optimal transport from unseen action prototypes to the set of all projected test videos. The resulting transport couplings in turn determine the target prototype for each unseen action. Rather than directly using the target prototype as final result, we re-position unseen action prototypes along the geodesic spanned by the original and target prototypes as a form of semantic regularization. For universal object models, we outline a variant that defines target prototypes based on an optimal transport between unseen action prototypes and object prototypes. Empirically, we show that universal prototype transport diminishes the biased selection of unseen action prototypes and boosts both universal action and object models for zero-shot classification and spatio-temporal localization.

AN END-TO-END DEEP LEARNING WORKFLOW FOR BUILDING SEGMENTATION, BOUNDARY REGULARIZATION AND VECTORIZATION OF BUILDING FOOTPRINTS

Abstract. Automatic building footprint extraction from remote sensing imagery is a widely used method, with deep learning techniques being particularly effective. However, deep learning approaches still require additional post-processing steps due to pixel-wise predictions, that contribute to occluded and geometrically incorrectly segmented buildings. To address this issue, we propose an end-to-end workflow that utilizes binary semantic segmentation, regularization, and vectorization. We implement and assess the performance of four convolutional neural network architectures including U-Net, U-NetFormer, FT-UnetFormer, and DCSwin on the MapAI Precision in Building Segmentation competition. To additionally improve the shape of the predicted buildings we apply regularization on the predictions to assess whether regularization further improves the geometrical shape and improve the prediction accuracy. We aim to produce accurate predictions with regularized boundaries that can prove useful in many cartographic and engineering applications. The regularization and vectorization workflow is further developed into a working QGIS-plugin that can be used to extend the functionality of QGIS. Our aim is to provide an end-to-end workflow for building segmentation, regularization and vectorization.

Domain generalization via adversarial out-domain augmentation for remaining useful life prediction of bearings under unseen conditions

Since classical deep learning (DL) techniques are hungry for massive data and suffer from domain shift, domain adaptation (DA) methods are broadly adopted in prognostics and health management (PHM) to align source and target domains. However, DA relies on target datasets collected in advance, which are not always available in practice. In this paper, a domain generalization (DG) approach, which learns from multiple source domains and generalizes well to unseen domains, is introduced for remaining useful life (RUL) prediction of bearings under unseen operating conditions. Specifically, we propose an adversarial out-domain augmentation (AOA) framework to generate pseudo-domains, thereby increasing the diversity of available samples. Hence, a generator is trained in an adversarial manner to generate augmented pseudo-domains by maximizing the domain discrepancy of the latent representations. In addition, we add manifold and semantic regularization to its objective function to ensure the consistency of the pseudo-domains. Trained with these available domains, a task predictor can improve the generalization in inaccessible target domain. Based on this, we provide a specific implementation of AOA-based RUL prediction for DG and validate its effectiveness and superiority using experimental datasets.

Binary multi-modal matrix factorization for fast item cold-start recommendation

Hashing technology can support large-scale recommendation very effectively due to its advantages of low storage cost and high recommendation efficiency. However, existing hashing-based recommendation methods often suffer from item cold-start problem. This is because they simply consider the user-item interaction and the single content information of the items, but the full interaction history is not always available and the single auxiliary information may be missing. To solve this issue, in this paper we propose a Binary Multi-modal Matrix Factorization (BMMF) method. First, we propose an efficient consensus multi-modal mapping to transform the heterogeneous multi-modal features to the unified factors by exploiting the complementarity of multiple modalities. Then, binary matrix factorization is simultaneously performed on the multi-modal features of the items and past user preferences to learn the compact binary codes of the users/items in a common Hamming space. In addition, inspired by the observation that similar instances often have similar binary codes within a short Hamming distance, we formulate a semantic structure regularization term to preserve the similarities of the items during the binary embedding process. Finally, we develop an effective Discrete Coordinate Descent (DCD) approach to tackle the formulated discrete hash optimization problem directly. Experiments on three publicly available real-world datasets demonstrate the superiority of the proposed method against the state-of-the-art methods. Our source codes and testing datasets are available athttps://github.com/pcm1217/BMMF.

Activation Modulation and Recalibration Scheme for Weakly Supervised Semantic Segmentation

Image-level weakly supervised semantic segmentation (WSSS) is a fundamental yet challenging computer vision task facilitating scene understanding and automatic driving. Most existing methods resort to classification-based Class Activation Maps (CAMs) to play as the initial pseudo labels, which tend to focus on the discriminative image regions and lack customized characteristics for the segmentation task. To alleviate this issue, we propose a novel activation modulation and recalibration (AMR) scheme, which leverages a spotlight branch and a compensation branch to obtain weighted CAMs that can provide recalibration supervision and task-specific concepts. Specifically, an attention modulation module (AMM) is employed to rearrange the distribution of feature importance from the channel-spatial sequential perspective, which helps to explicitly model channel-wise interdependencies and spatial encodings to adaptively modulate segmentation-oriented activation responses. Furthermore, we introduce a cross pseudo supervision for dual branches, which can be regarded as a semantic similar regularization to mutually refine two branches. Extensive experiments show that AMR establishes a new state-of-the-art performance on the PASCAL VOC 2012 dataset, surpassing not only current methods trained with the image-level of supervision but also some methods relying on stronger supervision, such as saliency label. Experiments also reveal that our scheme is plug-and-play and can be incorporated with other approaches to boost their performance. Our code is available at: https://github.com/jieqin-ai/AMR.

CyCMIS: Cycle-consistent Cross-domain Medical Image Segmentation via diverse image augmentation.

Domain shift, a phenomenon when there exists distribution discrepancy between training dataset (source domain) and test dataset (target domain), is very common in practical applications and may cause significant performance degradation, which hinders the effective deployment of deep learning models to clinical settings. Adaptation algorithms to improve the model generalizability from source domain to target domain has significant practical value. In this paper, we investigate unsupervised domain adaptation (UDA) technique to train a cross-domain segmentation method which is robust to domain shift, and which does not require any annotations on the test domain. To this end, we propose Cycle-consistent Cross-domain Medical Image Segmentation, referred as CyCMIS, integrating online diverse image translation via disentangled representation learning and semantic consistency regularization into one network. Different from learning one-to-one mapping, our method characterizes the complex relationship between domains as many-to-many mapping. A novel diverse inter-domain semantic consistency loss is then proposed to regularize the cross-domain segmentation process. We additionally introduce an intra-domain semantic consistency loss to encourage the segmentation consistency between the original input and the image after cross-cycle reconstruction. We conduct comprehensive experiments on two publicly available datasets to evaluate the effectiveness of the proposed method. Results demonstrate the efficacy of the present approach.

Fusing image feature and title semantic regularization with adversarial networks for cross-modal retrieval

With the improvement of living standard, people pay more and more attention to their health. Food is the foundation of human life, modern society places more emphasis on a balanced diet, which controls fat, protein, carbohydrate and vitamin intake. So food computing is more and more important. Food retrieval is one of the important research directions in food computing. On the basis of food retrieval, we can predict ingredients and instructions in each dish, according to the ingredients and instructions to speculate on the visual effect of cooking, which can guide human reasonable diet, analyse human diet structure and diet culture and so on. In this paper, we focus on cross-modal retrieval between food image and recipe. Firstly, we analyze the problems existing in the present method. Based on the problems existing in the existing method, a fusion image feature and title regularization with adversarial network is proposed, which uses the idea of generative adversarial to align the modes, fuses the local features and global features of the image, and adds the semantic regularity of title to improve the accuracy of the retrieval.

Intra-class low-rank regularization for supervised and semi-supervised cross-modal retrieval

Cross-modal retrieval aims to retrieve related items across different modalities, for example, using an image query to retrieve related text. The existing deep methods ignore both the intra-modal and inter-modal intra-class low-rank structures when fusing various modalities, which decreases the retrieval performance. In this paper, two deep models (denoted as ILCMR and Semi-ILCMR) based on intra-class low-rank regularization are proposed for supervised and semi-supervised cross-modal retrieval, respectively. Specifically, ILCMR integrates the image network and text network into a unified framework to learn a common feature space by imposing three regularization terms to fuse the cross-modal data. First, to align them in the label space, we utilize semantic consistency regularization to convert the data representations to probability distributions over the classes. Second, we introduce an intra-modal low-rank regularization, which encourages the intra-class samples that originate from the same space to be more relevant in the common feature space. Third, an inter-modal low-rank regularization is applied to reduce the cross-modal discrepancy. To enable the low-rank regularization to be optimized using automatic gradients during network back-propagation, we propose the rank-r approximation and specify the explicit gradients for theoretical completeness. In addition to the three regularization terms that rely on label information incorporated by ILCMR, we propose Semi-ILCMR in the semi-supervised regime, which introduces a low-rank constraint before projecting the general representations into the common feature space. Extensive experiments on four public cross-modal datasets demonstrate the superiority of ILCMR and Semi-ILCMR over other state-of-the-art methods.

Adaptive Framework for Deep Learning Based Dynamic and Temporal Topic Modeling from Big Data

Background: The large amount of data emanated from social media platforms need scalable topic modeling in order to get current trends and themes of events discussed on such platforms. Topic modeling play crucial role in many natural language processing applications like sentiment analysis, recommendation systems, event tracking, summarization, etc. Objective: The aim of the proposed work is to adaptively extract the dynamically evolving topics over streaming data, and infer the current trends and get the notion of trend of topics over time. Because of various world level events, many uncorrelated streaming channels tend to start discussion on similar topics. We aim to find the effect of uncorrelated streaming channels on topic modeling when they tend to start discussion on similar topics. Methods: An adaptive framework for dynamic and temporal topic modeling using deep learning has been put forth in this paper. The framework approximates online latent semantic indexing constrained by regularization on streaming data using adaptive learning method. The framework is designed using deep layers of feedforward neural network. Results: This framework supports dynamic and temporal topic modeling. The proposed approach is scalable to large collection of data. We have performed exploratory data analysis and correspondence analysis on real world Twitter dataset. Results state that our approach works well to extract topic topics associated with a given hashtag. Given the query, the approach is able to extract both implicit and explicit topics associated with the terms mentioned in the query. Conclusion: The proposed approach is a suitable solution for performing topic modeling over Big Data. We are approximating the Latent Semantic Indexing model with regularization using deep learning with differentiable ℓ1 regularization, which makes the model work on streaming data adaptively at real-time. The model also supports the extraction of aspects from sentences based on interrelation of topics and thus, supports aspect modeling in aspect-based sentiment analysis.

Intensity-Aware Single-Image Deraining With Semantic and Color Regularization.

Rain degrades image visual quality and disrupts object structures, obscuring their details and erasing their colors. Existing deraining methods are primarily based on modeling either visual appearances of rain or its physical characteristics (e.g., rain direction and density), and thus suffer from two common problems. First, due to the stochastic nature of rain, they tend to fail in recognizing rain streaks correctly, and wrongly remove image structures and details. Second, they fail to recover the image colors erased by heavy rain. In this paper, we address these two problems with the following three contributions. First, we propose a novel PHP block to aggregate comprehensive spatial and hierarchical information for removing rain streaks of different sizes. Second, we propose a novel network to first remove rain streaks, then recover objects structures/colors, and finally enhance details. Third, to train the network, we prepare a new dataset, and propose a novel loss function to introduce semantic and color regularization for deraining. Extensive experiments demonstrate the superiority of the proposed method over state-of-the-art deraining methods on both synthesized and real-world data, in terms of visual quality, quantitative accuracy, and running speed.