Target Image Research Articles

Improved Algorithm for Vehicle Bottom Safety Detection Based on YOLOv8n: PSP-YOLO

The safety inspection system for the underside of vehicles demands both high detection speed and accuracy, necessitating a model with small parameters and high network efficiency. To address these issues and improve the real-time performance of under-vehicle safety inspection, the YOLOv8n object-detection algorithm was enhanced, resulting in the PSP-YOLO algorithm. Firstly, the improved PSP-YOLO network introduces a P2 small-object detection head, which is particularly effective in detecting tiny objects, such as small dangerous objects under vehicles, thereby significantly enhancing the ability to detect such small targets. Secondly, the Space-to-Depth Conv (SPD-Conv) module was introduced into the backbone network, which improves the detection performance of low-quality samples and small objects while ensuring high precision and accurate localization of small targets in images. Finally, the Partial Convolution (PConv) lightweight module was also added to the feature fusion network, effectively reducing the model’s parameters and computational load, conserving computing resources, and improving detection speed and accuracy. Experimental results demonstrate that the modified model increased mAP@0.5 by 3.9% and mAP@0.5-0.95 by 3.6% on the dataset used, reduced the number of parameters to 2.53 M, and achieved a detection speed of 104.6 f/s. Therefore, the improved YOLOv8n algorithm significantly enhances the detection performance of dangerous objects, meeting the high-precision and real-time requirements for vehicle underside safety inspections.

Slow-Moving Ground Target Imaging Using Vortex Synthetic Aperture Radar

Abstract A vortex synthetic aperture radar (SAR) can obtain more target information when combined with orbital angular momentum (OAM). However, ground-moving targets can cause image defocusing owing to their azimuthal velocity component and imaging position offset issues due to their range velocity component. However, more information on moving targets can be acquired for analysis using vortex SAR. This study first established an imaging model to extract moving target information from vortex echo signals. The target Doppler parameters were then estimated, and an improved range-Doppler algorithm was applied to compensate for the Bessel term and azimuthal phase term. A new azimuth-matched filter was designed based on the Doppler parameters to achieve focused imaging of moving targets. The simulation results verified the effectiveness of the proposed algorithm, and its performance was compared with that of plane-wave SAR. Vortex SAR offered better target quality parameters and azimuth resolution for the same synthetic aperture length. The proposed algorithm can effectively improve the azimuthal imaging of moving targets in low-tomoderate OAM modes.

High-Precision ISAR Imaging of Translational Target Using Subaperture Segmentation and Motion Reconstruction

High-Precision ISAR Imaging of Translational Target Using Subaperture Segmentation and Motion Reconstruction

Ultrahigh-performance and broadband photodetector from visible to shortwave infrared band based on GaAsSb nanowires

Novel compact multispectral photodetectors detect various wavelengths, enabling visible light (VIS) and short-wave infrared (SWIR) dual-band detection for target identification and imaging. Nevertheless, combining VIS and SWIR photodetectors for multispectral detection faces challenges due to structural design and integration complexities. To streamline integration hurdles, high-Sb composition large-diameter GaAsSb nanowire photodetectors were developed, minimizing recombination centers and boosting gain. These photodetectors exhibit a high response in the light range from VIS (400 nm) to SWIR (2000 nm), exhibiting an ultrasensitive rise time (τr) of ∼0.8 ms, ultrahigh responsivity (R) of ∼2.32 × 105 A/W, and excellent detectivity (D*) of ∼9.10 × 1013 Jones in VIS (532 nm), and rapid τr of ∼6.5 ms, high R of ∼4.83 × 103 A/W with D* of ∼6.43 × 1011 Jones in SWIR (1300 nm), surpassing most previously reported III-V NW-based photodetectors. The simplified single GaAsSb NW photodetector significantly reduces the structural complexity and integration, achieving an ultrahigh-performance broadband optical response from the VIS to the SWIR region. The proposed broadband photodetector offers a potential approach for streamlining the integration of VIS/SWIR technologies and is expected to be applied in optical communication, monitoring, and imaging systems.

Perceptual Video Coding for Machines via Satisfied Machine Ratio Modeling.

Video Coding for Machines (VCM) aims to compress visual signals for machine analysis. However, existing methods only consider a few machines, neglecting the majority. Moreover, the machine's perceptual characteristics are not leveraged effectively, resulting in suboptimal compression efficiency. To overcome these limitations, this paper introduces Satisfied Machine Ratio (SMR), a metric that statistically evaluates the perceptual quality of compressed images and videos for machines by aggregating satisfaction scores from them. Each score is derived from machine perceptual differences between original and compressed images. Targeting image classification and object detection tasks, we build two representative machine libraries for SMR annotation and create a large-scale SMR dataset to facilitate SMR studies. We then propose an SMR prediction model based on the correlation between deep feature differences and SMR. Furthermore, we introduce an auxiliary task to increase the prediction accuracy by predicting the SMR difference between two images in different quality. Extensive experiments demonstrate that SMR models significantly improve compression performance for machines and exhibit robust generalizability on unseen machines, codecs, datasets, and frame types.

CoVR-2: Automatic Data Construction for Composed Video Retrieval.

Composed Image Retrieval (CoIR) has recently gained popularity as a task that considers both text and image queries together, to search for relevant images in a database. Most CoIR approaches require manually annotated datasets, comprising image-text-image triplets, where the text describes a modification from the query image to the target image. However, manual curation of CoIR triplets is expensive and prevents scalability. In this work, we instead propose a scalable automatic dataset creation methodology that generates triplets given video-caption pairs, while also expanding the scope of the task to include Composed Video Retrieval (CoVR). To this end, we mine paired videos with a similar caption from a large database, and leverage a large language model to generate the corresponding modification text. Applying this methodology to the extensive WebVid2M collection, we automatically construct our WebVid-CoVR dataset, resulting in 1.6 million triplets. Moreover, we introduce a new benchmark for CoVR with a manually annotated evaluation set, along with baseline results. We further validate that our methodology is equally applicable to image-caption pairs, by generating 3.3 million CoIR training triplets using the Conceptual Captions dataset. Our model builds on BLIP-2 pretraining, adapting it to composed video (or image) retrieval, and incorporates an additional caption retrieval loss to exploit extra supervision beyond the triplet, which is possible since captions are readily available for our training data by design. We provide extensive ablations to analyze the design choices on our new CoVR benchmark. Our experiments also demonstrate that training a CoVR model on our datasets effectively transfers to CoIR, leading to improved state-of-the-art performance in the zero-shot setup on the CIRR, FashionIQ, and CIRCO benchmarks.

3D human model guided pose transfer via progressive flow prediction network

Human pose transfer is to transfer a conditional person image to a new target pose. The difficulty lies in modeling the large-scale spatial deformation from the conditional pose to the target one. However, the commonly used 2D data representations and one-step flow prediction scheme lead to unreliable deformation prediction because of the lack of 3D information guidance and the great changes in the pose transfer. Therefore, to bring the original 3D motion information into human pose transfer, we propose to simulate the generation process of real person image. We drive the 3D human model reconstructed from the conditional person image with the target pose and project it to the 2D plane. The 2D projection thereby inherits the 3D information of the poses which can guide the flow prediction. Furthermore, we propose a progressive flow prediction network consisting of two streams. One stream is to predict the flow by decomposing the complex pose transformation into multiple sub-transformations. The other is to generate the features of the target image according to the predicted flow. Besides, to enhance the reliability of the generated invisible regions, we use the target pose information which contains structural information from the flow prediction stream as the supplementary information to the feature generation. The synthesized images with accurate depth information and sharp details demonstrate the effectiveness of the proposed method.

End to end polysemantic cooperative mixed task trainer for UAV target detection

With the rapid advancement and application of Unmanned Aerial Vehicles (UAVs), target detection in urban scenes has made significant progress. Achieving precise 3D reconstruction from oblique imagery is essential for accurate urban object detection in UAV images. However, challenges persist due to low detection accuracy caused by subtle target features, complex backgrounds, and the prevalence of small targets. To address these issues, we introduce the Polysemantic Cooperative Detection Transformer (Pc-DETR), a novel end-to-end UAV image target detection network. Our primary innovation, the Polysemantic Transformer (PoT) Backbone, enhances visual representation by leveraging contextual information to guide a dynamic attention matrix. This matrix, formed through convolutions, captures both static and dynamic features, resulting in superior detection. Additionally, we propose the Polysemantic Cooperative Mixed-Task Training scheme, which employs multiple auxiliary heads for diverse label assignments, boosting the encoder’s learning capacity. This approach customizes queries and optimizes training efficiency without increasing inference costs. Comparative experiments show that Pc-DETR achieves a 3% improvement in detection accuracy over the current state-of-the-art MFEFNet, setting a new benchmark in UAV image detection and advancing methodologies for intelligent UAV surveillance systems.

Unveiling the Structure-Fluorogenic Property Relationship of Seoul-Fluor-Derived Bioorthogonal Tetrazine Probes.

Tetrazine (Tz)-embedded fluorescent probes, known for their fluorogenicity following bioorthogonal inverse electron-demand Diels-Alder (iEDDA) reactions, are extensively used in bioimaging. Despite significant research on fluorogenic Tz probes, there has been limited systematic exploration of their fluorogenic responses with various dienophiles. In this study, we elucidate the structure-fluorogenic property relationship of bioorthogonal Tz probes. We synthesized a series of Seoul-Fluor-Tz (SFTz) probes designed to exhibit differentiated turn-on fluorescence upon iEDDA reactions with three dienophiles: trans-cyclooctene (TCO), bicyclo[6.1.0]nonyne (BCN), and spiro[2.3]hex-1-ene (Sph). Our findings revealed that the fluorogenic properties of the SFTz probes are highly dependent on the structures of dienophile-Tz adducts. By systematically modifying the electronic properties and employing quantum chemical calculations, we developed a series of SFTz probes with optimal dienophile-dependent fluorescence. These probes enabled simultaneous dual-color imaging of different cellular targets using a single probe, providing a robust approach for advanced bioimaging applications that require precise and efficient multicolor labeling strategies.

Closing the gap in domain adaptation for semantic segmentation: a time-aware method

Semantic segmentation models need a large number of images to be effectively trained but manual annotation of such images has a high cost. Active domain adaptation addresses this problem by pretraining the model with a synthetically generated dataset and then fine-tuning it with a few selected label annotations (the “budget”) on real images to account for the domain shift. Previous works annotate a percentage of either individual pixels or whole target images. We argue that the first is infeasible in practice, and the second spends part of the budget on classes that the pretrained model may have already learned well. We propose a method based on the annotation of regions computed by Segment Anything, a recently introduced foundation model for class-agnostic image segmentation. The key idea is to assign a ground truth label to each of a tiny subset of regions, those for which the model is more uncertain. In order to increase the number of annotated regions we propagate the ground truth labels to most similar regions according to a hierarchical clustering algorithm that uses the features learned by the pretrained model. Our method outperforms the state-of-the-art on the GTA5 to Cityscapes benchmark by using fewer annotations, almost closing the gap between the synthetically pre-trained model and that obtained with full supervision of the real images. Furthermore, we present competitive results for budgets less than 1% of samples and also for a larger and more challenging target dataset, Mapillary Vistas.

Late-stage (radio)fluorination of alkyl phosphonates via electrophilic activation

Constructing organic fluorophosphines, vital drug skeletons, through the direct fluorination of readily available alkyl phosphonates has been impeded due to the intrinsic low electrophilicity of PV and the high bond energy of P═O bond. Here, alkyl phosphonates are electrophilically activated with triflic anhydride and N-heteroaromatic bases, enabling nucleophilic fluorination at room temperature to form fluorophosphines via reactive phosphine intermediates. This approach facilitates the late-stage (radio)fluorination of broad dialkyl and monoalkyl phosphonates. Monoalkyl phosphonates derived from targeted drugs, including cyclophosphamide, vortioxetine, and dihydrocholesterol, are effectively fluorinated, achieving notable yields of 47−71%. Radiofluorination of medically significant 18F-tracers and synthons are completed in radiochemical conversions (radio-TLC) of 51−88% and molar activities up to 251 ± 12 GBq/μmol (initial activity 11.2 GBq) within 10 min at room temperature. Utilizing a phosphonamidic fluoride building block (BFPA), [18F]BFPA-Flurpiridaz and [18F]BFPA-E[c(RGDyK)]2 demonstrate high-contrast target imaging, excellent pharmacokinetics, and negligible defluorination.

Achieving High-Accuracy Target Recognition Using Few ISAR Images via Multi-Prototype Network with Attention Mechanism

Inverse synthetic aperture radar (ISAR) is a significant means of detection in space of non-cooperative targets, which means that the imaging geometry and associated parameters between the ISAR platform and the detection targets are unknown. In this way, a large number of ISAR images for high-accuracy target recognition are difficult to obtain. Recently, prototypical networks (PNs) have gained considerable attention as an effective method for few-shot learning. However, due to the specificity of the ISAR imaging mechanism, ISAR images often have unknown range and azimuth distortions, resulting in a poor imaging effect. Therefore, this condition poses a challenge for a PN to represent a class through a prototype. To address this issue, we use a multi-prototype network (MPN) with attention mechanism for ISAR image target recognition. The use of multiple prototypes eases the uncertainty associated with the fixed structure of a single prototype, enabling the capture of more comprehensive target information. Furthermore, to maximize the feature extraction capability of MPN for ISAR images, this method introduces the classical convolutional block attention module (CBAM) attentional mechanism, where CBAM generates attentional feature maps along channel and spatial dimensions to generate multiple robust prototypes. Experimental results demonstrate that this method outperforms state-of-the-art few-shot methods. In a four-class classification task, it achieved a target recognition accuracy of 95.08%, representing an improvement of 9.94–17.49% over several other few-shot approaches.

UUVDNet: An efficient unmanned underwater vehicle target detection network for multibeam forward-looking sonar

Precise localization of unmanned underwater vehicle (UUV) on multi-beam forward-looking sonar (MFLS) is a key technology in underwater robotic exploration. However, large appearance change and weak features of targets in MFLS image, pose a serious challenge on target detection. This paper proposes an efficient UUV detection model with an enhanced training scheme and convolutional block attention mechanism. The proposed training scheme is composed of dual-path gradient information optimization and adaptive loss adjustment, which improves the network’s detection capabilities and the underwater environment adaptability. The convolutional block attention module (CBAM) is developed for extracting channel and spatially weighted feature maps, and providing a mechanism to adaptive feature refinement. We construct a comprehensive sonar detection dataset from field experiments. By comparing existing state-of-the-art detection algorithms, our proposed UUVDNet outperforms Faster R-CNN, Cascade R-CNN, CenterNet, SSD, YOLOv8n, DPFIN, and MBSNN in terms of mAP50–95 by 39.5%, 35.7%, 8.4%, 14.0%, 2.5%, 13.6%, and 21.8%, respectively. Furthermore, our inference model size is the most compact among all detection models, with a size of 5.1M. Final ablation experiments effectively showcase the proposed components’ efficiency.

Multivariate Pattern Analysis of EEG Reveals Neural Mechanism of Naturalistic Target Processing in Attentional Blink.

The human brain has inherent limitations in consciously processing visual information. When individuals monitor a rapid sequence of images for detecting two targets, they often miss the second target (T2) if it appears within a short time frame of 200-500ms after the first target (T1), a phenomenon known as the attentional blink (AB). The neural mechanism behind the attentional blink (AB) remains unclear, largely due to the use of simplistic visual items such as letters and digits in conventional AB experiments, which differ significantly from naturalistic vision. This study employs advanced multivariate pattern analysis (MVPA) of human EEG data (including 17 females and 18 males) to explore the neural representations associated with target processing within a naturalistic paradigm under conditions where AB does or does not occur. Our MVPA analysis successfully decoded the identity of target images from EEG data. Moreover, in the AB condition, characterized by a limited time between targets, Tl processing coincided with T2 processing, resulting in the suppression of late representational markers of both Tl and T2. Conversely, in the condition with longer inter-target interval, neural representations endured for a longer duration. These findings suggest that the attentional blink can be attributed to the suppression of neural representations in the later stages of target processing.Significance Statement Within a naturalistic paradigm, we investigated the phenomenon known as attentional blink, where individuals struggle to identify a second target in a rapid sequence when the first target precedes it too closely. Attentional blink is purported to reflect an apparent bottleneck in the attention system's ability to rapidly redirect attentional resources; however, the mechanism underlying this phenomenon remains hotly debated. Our findings reveal that during a rapid presentation of natural images, a short temporal gap between targets results in reduced neural repre-sentations of targets and the occurrence of attentional blink. Conversely, when a greater temporal gap exists between targets, neural representations are preserved. This study provides valuable insights into how the human brain per-ceives the ever-changing visual world around us.

Enhanced multi-scale trademark element detection using the improved DETR

The exponential growth in the number of registered trademarks, coupled with the escalating incidents of trademark infringement, has made the automatic detection of such infractions a crucial area of study in the domain of market regulation. In light of the diverse range of elements and the pervasive presence of small targets in trademark images, we present an enhanced version of the DETR-based Multi-Scale Trademark Element Detection Network (MSTED-Net). Our primary innovation lies in incorporating a dual fusion mechanism that integrates the Spatial Attention Module (SAM) and Global Context Network (GCNet) within the backbone network, thereby providing a more robust approach to capture the essential characteristics of the trademark images under investigation. Subsequently, we develop a Multi-scale Feature Augmentation Pyramid (MFA-FPN), which aims to further fortify the model’s ability to extract features and boost the detection efficiency for small targets. The efficacy of our proposed detection network is demonstrated through experimental results, showcasing an outstanding detection accuracy of 91.12% in comparison to other state-of-the-art detection algorithms.

Research on Video Monitoring Technology for Galloping of OCS Additional Conductors of High-Speed Railway in Strong Wind Zone

The strong wind environment causes the additional conductor of the overhead contact system (OCS) of the Lanzhou–Xinjiang high-speed railway to gallop, significantly impacting the safe operation of the train. This paper presents the design of an online monitoring system for the galloping of additional conductors in the OCS, utilizing video monitoring for accurate and real-time assessment. Initially, the dynamics of the OCS additional conductor and its operational environment are examined, leading to the selection of suitable data transmission and power supply methods to finalize the camera configuration. Secondly, a preprocessing method for enhancing images of galloping in OCS additional conductors is developed, effectively reducing noise in edge detection through a region chain code clustering analysis. The video monitoring system effectively extracts wire edges, addressing the issues of splitting, breakage, and edge overlap in edge detection, while accurately identifying wire targets in video images. In conclusion, a galloping monitoring test platform is established to extract galloping data from additional conductors through video monitoring. The analysis of the galloping frequency and amplitude facilitates the comprehensive monitoring and assessment of the galloping status of OCS additional conductors. The video monitoring system effectively extracts and analyzes galloping data of the OCS additional conductor, fulfilling the fundamental requirements for the online monitoring of additional conductor galloping, and possesses significant engineering application value.

An Improved YOLOv8 Network for Multi-Object Detection with Large-Scale Differences in Remote Sensing Images

Aiming to address the challenges of low object detection precision in remote sensing images due to high background complexity and significant target scale variations, a novel model for large-scale disparate object detection in remote sensing images is proposed based on the modified YOLOv8. The model incorporates a Context Aggregation Module (CAM) with an attention mechanism in the backbone network to exploit contextual information, enabling multi-scale feature fusion for effective small object detection. The Neck network utilizes GSConv modules and employs the Slim-Neck design paradigm to enhance model robustness, making it better suited for detecting objects in high-complexity backgrounds. Furthermore, the model adopts the Wise-IoU as the loss function, incorporating a dynamic nonmonotonic focusing mechanism and a gradient gain allocation strategy to enhance the overall performance of disparate object detection. The experimental results indicate that promising performance improvements in the face of large-scale variations in remote sensing image targets. Specifically, the model achieves mAP values of 71.4% and 90.3% on the DOTA and NWPU VHR-10 datasets, respectively, representing increases of 4.4% and 3.7% compared to the original model. Compared to other typical algorithms, it also has considerable advantages in both comprehensive detection accuracy and detection speed.

Satellite Image Small Target Super-Resolution Detection Based on ACGAN and Enhanced DETR

In recent years, with the rapid development of deep learning, neural network and other related technologies and their wide application in image processing research, the degree of intelligence in the field of image processing research is getting higher and higher. Especially in the research of image small target detection, because of its relatively small shape, image small target has relatively few and unclear feature information and detail information, which makes it easy to make errors and omissions in recognition and detection. To improve the precision of satellite image small target detection, this paper processes a small image target super-resolution detection based on ACGAN and enhanced DETR. In this paper, ACGAN is used to classify and generate the feature information of small targets in images to enhance the feature information of small targets to be measured. Second, the DETR algorithm model is improved, and a small target detection algorithm model with image super-resolution enhancement is designed to generate high-resolution images by super-resolution reconstruction of images with enhanced target feature information to improve the detection accuracy of image small targets. The experiments in this paper show that after the processing of target feature enhancement and image super-resolution quality enhancement, the detection accuracy of image small targets has been improved significantly.

Optimizing tomato detection and counting in smart greenhouses: A lightweight YOLOv8 model incorporating high- and low-frequency feature transformer structures

ABSTRACT Tomato harvesting in intelligent greenhouses is crucial for reducing costs and optimizing management. Agricultural robots, as an automated solution, require advanced visual perception. This study proposes a tomato detection and counting algorithm based on YOLOv8 (TCAttn-YOLOv8). To handle small, occluded tomato targets in images, a new detection layer (NDL) is added to the Neck and Head decoupled structure, improving small object recognition. The ColBlock, a dual-branch structure leveraging Transformer advantages, enhances feature extraction and fusion, focusing on densely targeted regions and minimizing small object feature loss in complex backgrounds. C2fGhost and GhostConv are integrated into the Neck network to reduce model parameters and floating-point operations, improving feature expression. The WIoU (Wise-IoU) loss function is adopted to accelerate convergence and increase regression accuracy. Experimental results show that TCAttn-YOLOv8 achieves an mAP@0.5 of 96.31%, with an FPS of 95 and a parameter size of 2.7 M, outperforming seven lightweight YOLO algorithms. For automated tomato counting, the R2 between predicted and actual counts is 0.9282, indicating the algorithm’s suitability for replacing manual counting. This method effectively supports tomato detection and counting in intelligent greenhouses, offering valuable insights for robotic harvesting and yield estimation research.

TSFormer: Tracking Structure Transformer for Image Inpainting

Recent studies have shown that image structure can significantly facilitate image inpainting. However, current approaches mostly explore structure prior without considering its guidance to texture reconstruction, leading to performance degradation. To solve this issue, we propose a two-stream Tracking Structure Transformer (TSFormer), including structure target stream and image completion stream, to capture the synchronous and dynamic interplay between structure and texture. Specifically, we first design a structure enhancement module to restore the Histograms of Oriented Gradient (HOG) and the edge of an input image in a sketch space, which forms the input of the structure target stream. Meanwhile, in the image completion stream, we design a channel-space parallel-attention component to facilitate the efficient co-learning of channel and spatial visual cues. To build a bridge between the two streams, we further develop a structure-texture cross-attention module, wherein both structure and texture are synchronously extracted through self-attention, and texture extraction is implemented by dynamically tracking the structure in a cross-attention fashion, enabling the capture of the intricate interaction between structure and texture. Extensive experiments evaluated on three benchmark datasets, including CelebA, Places2, and Paris StreetView, demonstrate that the proposed TSFormer achieves state-of-the-art performance compared to its competitors. The code is available at https://github.com/GZHU-DVL/TSFormer .