Image Semantic Information Research Articles

Bi-Interfusion: A bidirectional cross-fusion framework with semantic-guided transformers in LiDAR-camera fusion

Multi-sensor modal fusion has shown significant advantages in 3D object detection tasks. However, existing methods for fusing multi-modal features into the bird’s eye view (BEV) space often encounter challenges such as feature misalignment, underutilization of semantic information, and inaccurate depth estimation on the Z-axis, resulting in suboptimal performance. To address these issues, we propose Bi-Interfusion, a novel multi-modal fusion framework based on transformers. Bi-Interfusion incorporates a bidirectional fusion architecture, including components such as Pixel-wise Semantic Painting, Gaussian Depth Prior Distribution module, and Semantic Guidance Align module, to overcome the limitations of traditional fusion methods. Specifically, Bi-Interfusion employs a bidirectional cross-fusion strategy to merge image and LiDAR features, enabling the generation of multi-sensor BEV features. This approach leverages a refined Gaussian Depth Prior Distribution generated from LiDAR points, thereby improving the precision of view transformation. Additionally, we apply a pixel-wise semantic painting technique to embed image semantic information into LiDAR point clouds, facilitating a more comprehensive scene understanding. Furthermore, a transformer-based model is utilized to establish soft correspondences among multi-sensor BEV features, capturing positional dependencies and fully exploiting semantic information for alignment. Through experiments on nuScenes benchmark dataset, Bi-Interfusion demonstrates notable performance improvements, achieving a competitive performance of 72.6% mAP and 75.4% NDS in the 3D object detection task.

DCSS-UNet: UNet based on State Space Model for Polyp Segmentation

Early and accurate segmentation of medical images can provide valuable information for medical treatment. In recent years, the automatic and accurate segmentation of polyps in colonoscopy images has received extensive attention from the research community of artificial intelligence and computer vision. Many researchers have conducted in-depth research on models based on CNN and Transformer. However, CNN have limited ability to model remote dependencies, which makes it challenging to fully utilize semantic information in images. On the other hand, the complexity of the secondary computation poses a challenge to the transformer. Recently, state-space models (SSMS), such as Mamba, have been recognized as a promising approach. They not only show superior performance in remote interaction, but also maintain linear computational complexity. Inspired by Mamba, we propose DCSS-UNet, where we utilize visual state space (VSS) blocks in VMamba to capture a wide range of contextual information. In the Skip connection phase, we propose Skip Connects Feature Attention modules(SFA) to better communicate information from the encoder. In the decoder stage, we innovatively combined the Temporal Fusion Attention Module(TFAM) to effectively fuse the feature information. In addition, we introduced a custom Loss calculation method, Tversky Loss, for the model to achieve faster convergence and improve segmentation along polyp boundaries. Our model was trained on the Kvasir-SEG and CVC-ClinicDB datasets, and validated on datasets Kvasir-SEG, CVC-ColonDB, CVC-300, and ETIS. The results show that the model achieves good segmentation accuracy and generalization performance with a low number of parameters. We are 6.1% ahead in the Kavirs-SEG dataset and 3.1% ahead in the CVC-ClinicDB dataset compared to VM-UNet.

Multimodal Recipe Recommendation with Heterogeneous Graph Neural Networks

Recipe recommendation is the process of recommending suitable recipes to users based on factors such as user preferences and dietary needs. Recipes typically involve multiple modalities, with text and images being common, while most typical recipe recommendation methods recommend recipes to users based on text. Obviously, the expressiveness of a single modal is often not enough, and the semantic information of images is more abundant. Moreover, it is difficult to grasp the feature fusion granularity of different kinds of modal information and the relationship between recipes and users. To solve the above problem, this paper proposes a Multimodal Heterogeneous Graph Neural Network Recipe Recommendation (MHGRR) architecture, which aims to fully fuse the various kinds of modal information of recipes and handle the relationship between users and recipes. We use embedding and shallow Convolutional Neural Networks (CNNs) to extract original text and image information for unifying feature fusion granularity, and use Heterogeneous Graph Neural Networks based on GraphSAGE to capture the complex relationship between users and recipes. To verify the effectiveness of our proposed model, we perform some comparative experiments on a real dataset; the experiments show that our method outperforms most popular recipe recommendation methods. Through an ablation experiment, we found that adding image information to recipe recommendation is more effective, and we additionally found that as the output dimensions of GraphSAGE increased, the performance of the model varied little.

SAM-RSP: A new few-shot segmentation method based on segment anything model and rough segmentation prompts

Few-shot segmentation (FSS) aims to segment novel classes with a few labeled images. The backbones used in existing methods are pre-trained through classification tasks on the ImageNet dataset. Although these backbones can effectively perceive the semantic categories of images, they cannot accurately perceive the regional boundaries within one image, which limits the model performance. Recently, Segment Anything Model (SAM) has achieved precise image segmentation based on point or box prompts, thanks to its excellent perception of region boundaries within one image. However, it cannot effectively provide semantic information of images. This paper proposes a new few-shot segmentation method that can effectively perceive both semantic categories and regional boundaries. This method first utilizes the SAM encoder to perceive regions and obtain the query embedding. Then the support and query images are input into a backbone pre-trained on ImageNet to perceive semantics and generate a rough segmentation prompt (RSP). This query embedding is combined with the prompt to generate a pixel-level query prototype, which can better match the query embedding. Finally, the query embedding, prompt, and prototype are combined and input into the designed multi-layer prompt transformer decoder, which is more efficient and lightweight, and can provide a more accurate segmentation result. In addition, other methods can be easily combined with our framework to improve their performance. Plenty of experiments on PASCAL-5i and COCO-20i under 1-shot and 5-shot settings prove the effectiveness of our method. Our method also achieves new state-of-the-art. Codes are available at https://github.com/Jiaguang-NEU/SAM-RSP.

HRDLNet: a semantic segmentation network with high resolution representation for urban street view images

Semantic segmentation of urban street scenes has attracted much attention in the field of autonomous driving, which not only helps vehicles perceive the environment in real time, but also significantly improves the decision-making ability of autonomous driving systems. However, most of the current methods based on Convolutional Neural Network (CNN) mainly use coding the input image to a low resolution and then try to recover the high resolution, which leads to problems such as loss of spatial information, accumulation of errors, and difficulty in dealing with large-scale changes. To address these problems, in this paper, we propose a new semantic segmentation network (HRDLNet) for urban street scene images with high-resolution representation, which improves the accuracy of segmentation by always maintaining a high-resolution representation of the image. Specifically, we propose a feature extraction module (FHR) with high-resolution representation, which efficiently handles multi-scale targets and high-resolution image information by efficiently fusing high-resolution information and multi-scale features. Secondly, we design a multi-scale feature extraction enhancement (MFE) module, which significantly expands the sensory field of the network, thus enhancing the ability to capture correlations between image details and global contextual information. In addition, we introduce a dual-attention mechanism module (CSD), which dynamically adjusts the network to more accurately capture subtle features and rich semantic information in images. We trained and evaluated HRDLNet on the Cityscapes Dataset and the PASCAL VOC 2012 Augmented Dataset, and verified the model’s excellent performance in the field of urban streetscape image segmentation. The unique advantages of our proposed HRDLNet in the field of semantic segmentation of urban streetscapes are also verified by comparing it with the state-of-the-art methods.

Multization: Multi-Modal Summarization Enhanced by Multi-Contextually Relevant and Irrelevant Attention Alignment

This article focuses on the task of Multi-Modal Summarization with Multi-Modal Output for China JD.COM e-commerce product description containing both source text and source images. In the context learning of multi-modal (text and image) input, there exists a semantic gap between text and image, especially in the cross-modal semantics of text and image. As a result, capturing shared cross-modal semantics earlier becomes crucial for multi-modal summarization. However, when generating the multi-modal summarization, based on the different contributions of input text and images, the relevance and irrelevance of multi-modal contexts to the target summary should be considered, so as to optimize the process of learning cross-modal context to guide the summary generation process and to emphasize the significant semantics within each modality. To address the aforementioned challenges, Multization has been proposed to enhance multi-modal semantic information by multi-contextually relevant and irrelevant attention alignment. Specifically, a Semantic Alignment Enhancement mechanism is employed to capture shared semantics between different modalities (text and image), so as to enhance the importance of crucial multi-modal information in the encoding stage. Additionally, the IR-Relevant Multi-Context Learning mechanism is utilized to observe the summary generation process from both relevant and irrelevant perspectives, so as to form a multi-modal context that incorporates both text and image semantic information. The experimental results in the China JD.COM e-commerce dataset demonstrate that the proposed Multization method effectively captures the shared semantics between the input source text and source images, and highlights essential semantics. It also successfully generates the multi-modal summary (including image and text) that comprehensively considers the semantics information of both text and image.

Cherry growth modeling based on Prior Distance Embedding contrastive learning: Pre-training, anomaly detection, semantic segmentation, and temporal modeling

In current plant phenotyping research, the study of plant time-series images based on deep learning has received widespread attention. While such image data is relatively easy to obtain, the cost of annotation is high. One efficient method for achieving cost-effective training is through contrastive learning. Plant growth is slow, and the changes in image sequences over a period of time are small, with simple semantic information. Previous contrastive pre-training models struggled to effectively distinguish positive samples from the same image with different augmented views and similar negative samples from different images. Therefore, this paper proposes a method called self-supervised contrastive learning method for plant time-series images with a Priori Distance Embedding (PDE). The semantic information in images corresponding to different phenological stages of plants varies. This method transforms this crucial domain knowledge into prior distances for image pairs and conducts contrastive learning pre-training. The learned weights can be transferred to downstream tasks. Building upon this method, experiments were conducted on cherry time-series images to assess the quality of pre-training through a plant phenotyping image semantic segmentation task. To provide a comprehensive example of plant time-series image phenotypic analysis, this paper establishes a cherry growth temporal model, specifically including PDE pre-training, anomaly detection, semantic segmentation, and recording the results from the temporal dimension. The experiments indicate that this self-supervised contrastive learning method can be effectively applied to the pre-training of plant time-series images, demonstrating broad applicability in various computer vision studies related to plant phenotyping.

LR3S: A lightweight semantic segmentation model for road scenes based on improved DeepLabV3+

In the field of autonomous driving, driving systems need to understand and quickly respond to changes in road scenes, which makes it equally important to enhance the accuracy and real-time performance of semantic segmentation tasks in road scenes. This article proposes a lightweight road scene semantic segmentation model LR3S that integrates global contextual information based on the DeepLabV3+ framework. LR3S utilizes a lightweight GhostNetV2 network as the backbone to capture rich semantic information in images, and uses ASPP_eSE module to enhance the capture of multi-scale and detail level semantic information. In addition, a lightweight CARAFE upsampling operator is utilized to upsample feature maps, taking advantage of CARAFE’s large receptive field and low computational cost to prevent the loss of fine-grained features and ensure the integrity of semantic information. Experimental results demonstrate that LR3S achieves an MIoU of 74.47% on the Cityscapes dataset and obtains an MIoU of 76.01% on the PASCAL VOC 2012 dataset. Compared to baseline semantic segmentation models, LR3S significantly reduces the parameter amount while maintaining segmentation accuracy, achieving a good balance between model accuracy and real-time performance.

Application of Image Segmentation Algorithms in Computer Vision

In the field of computer vision (CV), image segmentation technology, as a fundamental part, has a crucial impact on the accuracy of subsequent image processing tasks. Image segmentation is not only a crucial transitional step from image processing to image analysis, but also a hot and difficult research topic in the field of CV. Although significant progress has been made in the research of image segmentation algorithms, existing segmentation algorithms may still face challenges in certain specific scenarios due to the complexity and diversity of images, making it difficult to achieve ideal segmentation results. In recent years, the rapid development of deep learning (DL) technology has brought new breakthroughs to the field of image segmentation. DL models, especially Convolutional Neural Networks (CNNs), can capture semantic information of images more accurately by automatically learning feature representations in images, thereby achieving more precise image segmentation. This article delves into the research and application of image segmentation algorithms in CV, with a focus on the application of DL in the field of image segmentation. With the continuous development of advanced technologies such as DL, it is believed that image segmentation technology will play a greater role in more fields in the future.

A multi-label image classification method combining multi-stage image semantic information and label relevance

A multi-label image classification method combining multi-stage image semantic information and label relevance

Region-Focused Network for Dense Captioning

Dense captioning is a very critical but under-explored task, which aims to densely detect localized regions-of-interest (RoIs) and describe them with natural language in a given image. Although recent studies tried to fuse multi-scale features from different visual instances to generate more accurate descriptions, their methods still suffer from the lack of exploration of relation semantic information in images, leading to less informative descriptions. Furthermore, indiscriminately fusing all visual instance features will introduce redundant information, resulting in poor matching between descriptions and corresponding regions. In this work, we propose a Region-Focused Network (RFN) to address these issues. Specifically, to fully comprehend the images, we first extract the object-level features, and encode the interaction and position relations between objects to enhance the object representations. Then, to decrease the interference from redundant information about the target region, we extract the most relevant information to the region. Finally, a region-based Transformer is employed to compose and align the previous mined information and generate the corresponding descriptions. Extensive experiments on Visual Genome V1.0 and V1.2 datasets show that our RFN model outperforms the state-of-the-art methods, thus verifying its effectiveness. Our code is available at https://github.com/VILAN-Lab/DesCap .

Style-Enhanced Transformer for Image Captioning in Construction Scenes.

Image captioning is important for improving the intelligence of construction projects and assisting managers in mastering construction site activities. However, there are few image-captioning models for construction scenes at present, and the existing methods do not perform well in complex construction scenes. According to the characteristics of construction scenes, we label a text description dataset based on the MOCS dataset and propose a style-enhanced Transformer for image captioning in construction scenes, simply called SETCAP. Specifically, we extract the grid features using the Swin Transformer. Then, to enhance the style information, we not only use the grid features as the initial detail semantic features but also extract style information by style encoder. In addition, in the decoder, we integrate the style information into the text features. The interaction between the image semantic information and the text features is carried out to generate content-appropriate sentences word by word. Finally, we add the sentence style loss into the total loss function to make the style of generated sentences closer to the training set. The experimental results show that the proposed method achieves encouraging results on both the MSCOCO and the MOCS datasets. In particular, SETCAP outperforms state-of-the-art methods by 4.2% CIDEr scores on the MOCS dataset and 3.9% CIDEr scores on the MSCOCO dataset, respectively.

Multi-modal sarcasm detection based on Multi-Channel Enhanced Fusion model

The voluminous quantity of data accessible on social media platforms offers insight into the sentiment disposition of individual users, where multi-modal sarcasm detection is often confounding. Existing sarcasm detection methods use different information fusion methods to combine information from different modalities but ignore hidden information within modalities and inconsistent information between modalities. Discovering the implicit information within the modalities and strengthening the information interaction between modalities is still an important challenge. In this paper, we propose a Multi-Channel Enhanced Fusion (MCEF) model for cross-modal sarcasm detection to maximize the information extraction between different modalities. Specifically, text extracted from images acts as a new modality in the front-end fusion models to augment the utilization of image semantic information. Then, we propose a novel bipolar semantic attention mechanism to uncover the inconsistencies among different modal features. Furthermore, a decision-level fusion strategy from a new perspective is devised based on four models to achieve multi-channel fusion, each with a distinct focus, to leverage their advantages and mitigate the limitations. Extensive experiments demonstrate that our model surpasses current state-of-the-art models in multi-modal sarcasm detection.

Towards Unsupervised Referring Expression Comprehension with Visual Semantic Parsing

Referring Expression Comprehension (REC) is a task that involves grounding a specific object in an image based on a given referring query in the form of bounding boxes. Existing fully-supervised or weakly-supervised REC methods rely on either manually annotated regional coordinates or query texts. In this paper, we propose an unsupervised training paradigm for the REC task that does not require any manual annotated data. Specifically, we introduce a Visual-Semantic-Parsing-based Unsupervised Referring Expression Comprehension framework (VUREC), which leverages a Visual Semantic Parser (VSP) as its core module to recognize the rich semantic information in images and construct pseudo-region-query pairs as the training supervision for REC. The VSP comprises a Scene Graph Parser (SGP) and a Visual Concept Detector (VCD) that can detect the locations, categories, attributes of objects, and visual relationships among them in images. Furthermore, we present a Referring Expression Reasoning (RER) model that utilizes a Multi-Modal Cascade Attention Decoder (MCAD) for fine-grained multi-modality fusion and regresses the four coordinates of the referential object directly. The experimental results on three benchmark datasets of Refcoco, Refcoco+ and Refcocog demonstrate the effectiveness of our proposed method.

MTSTR: Multi-task learning for low-resolution scene text recognition via dual attention mechanism and its application in logistics industry.

Recognizing texts in images plays an important role in many applications, such as industrial intelligence, robot vision, automatic driving, command assistance, and scene understanding. Although great progress has been achieved in various fields, research on complex systems modeling using text recognition technology requires further attention. To address this, we propose a new end-to-end multi-task learning method, which includes a super-resolution branch (SRB) and a recognition branch. To effectively learn the semantic information of images, we utilize the feature pyramid network (FPN) to fuse high- and low-level semantic information. The feature map generated by FPN is then delivered separately to the super-resolution branch and the recognition branch. We introduce a novel super-resolution branch, the SRB based on the proposed dual attention mechanism (DAM), designed to enhance the capability of learning low-resolution text features. The DAM incorporates the residual channel attention to enhance channel dependencies and the character attention module to focus on context information. For the recognition branch, the feature map generated by FPN is fed into an RNN sequence module, and an attention-based decoder is constructed to predict the results. To address the issue of low-resolution text recognition in numerous Chinese scenes, we propose the Chinese super-resolution datasets instead of relying on traditional down-sampling techniques to generate training datasets. Experiments demonstrate that the proposed method performs robustly on low-resolution text images and achieves competitive results on benchmark datasets.

Improving the Accuracy of Robot Collecting Organisms in Marine Environment Based on Yolov5 Improvement

AbstractIn this paper, an improved YOLOv5 multiscale marine organism target detection algorithm (YOLOv5‐Mult) is proposed to address the insufficient feature extraction ability of small targets, low detection accuracy, and high catching error of existing models in complex environments. First, real frame clustering is performed using the Kmeans++ method. Second, the BiFPN network module is adopted in lieu of the PANet network module to enhance the feature fusion ability. Next, the multilayer semantic fusion module RBC (RepBlock CSP) replaces the C3 module before the SPP layer of the Backbone network and the C3 module in the Neck layer to enrich the image semantic information. Finally, the multiscale feature fusion module MC (Mult Conv) replaces the last C3 module in the Backbone network to mitigate the semantic gap between different feature channel layers. Experimental results demonstrate that the improved algorithm attains a mAP value of 71.18%, which is 5.22% higher than that of the original YOLOv5 algorithm, providing accurate identification and fishing for underwater robots.

AFCANet: An adaptive feature concatenate attention network for multi-focus image fusion

For multi-focus image fusion, the existing deep learning based methods cannot effectively learn the texture features and semantic information of the source image to generate high-quality fused images. Thus, we develop a new adaptive feature concatenate attention network named AFCANet, which adaptively learns cross-layer features and retains the texture features and semantic information of images to generate visually appealing fully focused images. In AFCANet, the encoder-decoder network is used as the backbone network. In the unsupervised training stage, an adaptive cross-layer skip connection mode is designed, and a cross-layer adaptive coordinate attention module is built to acquire meaningful information from the image along with ignoring unimportant information to obtain a better image fusion effect. In addition, in the middle of the encoder-decoder network, we also introduce an effective channel attention module to fully learn the output of the encoder, and accelerate network convergence. In the inference stage, we apply the pixel-based spatial frequency fusion rules to fuse the adaptive features learned by the encoder, which can successfully combine the texture and semantic information of the image and produce a more precise decision map. Extensive experiments on public datasets and the HBU-CVMDSP dataset show that our AFCANet can effectively improve the accuracy of the decision map in the focus and defocus regions, as well as improve the ability to retain the abundant details and edge features of the source image.

AdvMask: A sparse adversarial attack-based data augmentation method for image classification

Data augmentation has been an essential technique for improving the generalization ability of deep neural networks in image classification tasks. However, intensive changes in appearance and different degrees of occlusion in images are the key factors that severely affect the generalization ability of image classification models. Therefore, in order to enhance the generalization performance and robustness of deep models, data augmentation approaches by providing models with more diverse training data in various scenarios are widely applied. Although many existing data augmentation methods simulate occlusion in the augmented images to enhance the generalization of models, these methods randomly delete some areas in images without considering the semantic information of images. In this work, we propose a novel data augmentation method named AdvMask for image classification based on sparse adversarial attack techniques. AdvMask first identifies the key points that have the greatest influence on the classification results via a proposed end-to-end sparse adversarial attack module. During the data augmentation process, AdvMask efficiently generates diverse augmented data with structured occlusions based on the key points. By doing so, AdvMask can force deep models to seek other relevant content while the most discriminative content is hidden. Extensive experimental results on various benchmark datasets and deep models demonstrate that our proposed method can effectively improve the generalization performance of deep models and significantly outperforms previous data augmentation methods. Code for reproducing our results is available at https://github.com/Jackbrocp/AdvMask.

A Semantic Information-Based Optimized vSLAM in Indoor Dynamic Environments

In unknown environments, mobile robots can use visual-based Simultaneous Localization and Mapping (vSLAM) to complete positioning tasks while building sparse feature maps and dense maps. However, the traditional vSLAM works in the hypothetical environment of static scenes and rarely considers the dynamic objects existing in the actual scenes. In addition, it is difficult for the robot to perform high-level semantic tasks due to its inability to obtain semantic information from sparse feature maps and dense maps. In order to improve the ability of environment perception and accuracy of mapping for mobile robots in dynamic indoor environments, we propose a semantic information-based optimized vSLAM algorithm. The optimized vSLAM algorithm adds the modules of dynamic region detection and semantic segmentation to ORB-SLAM2. First, a dynamic region detection module is added to the vision odometry. The dynamic region of the image is detected by combining single response matrix and dense optical flow method to improve the accuracy of pose estimation in dynamic environment. Secondly, the semantic segmentation of images is implemented based on BiSeNet V2 network. For the over-segmentation problem in semantic segmentation, a region growth algorithm combining depth information is proposed to optimize the 3D segmentation. In the process of map building, semantic information and dynamic regions are used to remove dynamic objects and build an indoor map containing semantic information. The system not only can effectively remove the effect of dynamic objects on the pose estimation, but also use the semantic information of images to build indoor maps containing semantic information. The proposed algorithm is evaluated and analyzed in TUM RGB-D dataset and real dynamic scenes. The results show that the accuracy of our algorithm outperforms that of ORB-SLAM2 and DS-SLAM in dynamic scenarios.

A DCT probability histogram-based ROI features for content-based natural and medical image retrieval applications

The effectiveness of a Content-based Image Retrieval (CBIR) system depends on selecting representative and salient visual features. Capturing the salient image features for forming a feature vector is an essential part of the CBIR process, as it can significantly affect retrieval efficiency. Most images contain a visually prominent region that perceives maximum image semantic information, known as Regions-of-Interest (ROIs). This paper suggests a new CBIR scheme that employs a Graph-based Visual Saliency (GBVS) map to extract the ROI of the image. Further, block-level Discrete Cosine Transforms (DCT) and subsequent histograms are computed for the DC and significant AC coefficients of the extracted ROI image. In addition, a low-dimensional feature vector is formed from the modified histograms. Subsequently, six statistical features from the background scene have also been computed for the final feature vector construction. As a result, the final feature vector contains ROI and background information in their proper proportion to improve retrieval performance. Moreover, the performance of the proposed CBIR scheme has been evaluated using one medical, three natural, two objects, one texture, and one natural heterogeneous image database. All these databases collectively contain 53732 different images from 200 image classes. The outcomes of the experiments demonstrate substantial improvement in the retrieval process with some existing related schemes. At the same time, it also shows that the performance of the proposed scheme is robust concerning all kinds of image databases.