Object Regions Research Articles

Domain Adaptive Object Detection via Balancing between Self-Training and Adversarial Learning.

Deep learning based object detectors struggle generalizing to a new target domain bearing significant variations in object and background. Most current methods align domains by using image or instance-level adversarial feature alignment. This often suffers due to unwanted background and lacks class-specific alignment. A straightforward approach to promote class-level alignment is to use high confidence predictions on unlabeled domain as pseudo-labels. These predictions are often noisy since model is poorly calibrated under domain shift. In this paper, we propose to leverage model's predictive uncertainty to strike the right balance between adversarial feature alignment and class-level alignment. We develop a technique to quantify predictive uncertainty on class assignments and bounding-box predictions. Model predictions with low uncertainty are used to generate pseudo-labels for self-training, whereas the ones with higher uncertainty are used to generate tiles for adversarial feature alignment. This synergy between tiling around uncertain object regions and generating pseudo-labels from highly certain object regions allows capturing both image and instance-level context during the model adaptation. We report thorough ablation study to reveal the impact of different components in our approach. Results on five diverse and challenging adaptation scenarios show that our approach outperforms existing state-of-the-art methods with noticeable margins.

Brain dynamics during mnemonic discrimination in preclinical and prodromal Alzheimer’s disease

AbstractBackgroundMnemonic discrimination of similar lures and targets is emerging as a promising marker of subtle memory decline in early Alzheimer’s disease (AD) due to its reliance on regions most affected by AD pathology. Tau pathology initially targets medial temporal lobe (MTL) regions in an anterior‐temporal (AT) brain network, which is preferentially involved in object processing, before it can be found in the posterior‐medial (PM) network associated with scene processing. Tests of object and scene mnemonic discrimination may therefore serve as behavioural indicators of the functional integrity of AT and PM regions, respectively. We investigated behavioural and neural changes in mnemonic discrimination across the AD spectrum.Method201 DELCODE cohort participants underwent fMRI scanning during an object‐scene mnemonic discrimination task at four time points from baseline to three‐year follow‐up. At baseline, participants were split by diagnostic group (cognitively unimpaired = CU, subjective cognitive decline = SCD, mild cognitive impairment = MCI, dementia of the Alzheimer’s type = DAT; Table 1). We identified brain regions showing lower activity to repeated targets but not similar lures, indicative of successful pattern separation of overlapping memory representations.ResultThese findings present cross‐sectional baseline data. Behavioural data showed a main effect of group but no group by category (objects, scenes) interaction. CUs and SCDs outperformed patients across domains and CUs scored higher than SCDs only in scene discrimination. Amyloid positive and negative CUs and SCDs did not differ in performance. Successful mnemonic discrimination engaged hippocampus, parahippocampal cortex, precuneus, angular gyrus, and inferior frontal regions for both objects and scenes. Correct object discrimination was additionally associated with amygdala and transentorhinal cortical activity. Reduced brain activation in regions of the anterior‐temporal and posterior‐medial network during successful object and scene mnemonic discrimination, respectively, was found in patient groups (MCI+DAT) compared to cognitively normal (CU+SCD) older adults.ConclusionOur findings lend further support to the utility of mnemonic discrimination as indicator of memory impairment in early AD. Object and scene tasks engage partially dissociable brain networks and individuals with cognitive impairment show reduced neural pattern separation effects. These findings lay the groundwork for future longitudinal analyses to track these changes based on clinical and biomarker status.

QDTrack: Quasi-Dense Similarity Learning for Appearance-Only Multiple Object Tracking.

Similarity learning has been recognized as a crucial step for object tracking. However, existing multiple object tracking methods only use sparse ground truth matching as the training objective, while ignoring the majority of the informative regions in images. In this paper, we present Quasi-Dense Similarity Learning, which densely samples hundreds of object regions on a pair of images for contrastive learning. We combine this similarity learning with multiple existing object detectors to build Quasi-Dense Tracking (QDTrack), which does not require displacement regression or motion priors. We find that the resulting distinctive feature space admits a simple nearest neighbor search at inference time for object association. In addition, we show that our similarity learning scheme is not limited to video data, but can learn effective instance similarity even from static input, enabling a competitive tracking performance without training on videos or using tracking supervision. We conduct extensive experiments on a wide variety of popular MOT benchmarks. We find that, despite its simplicity, QDTrack rivals the performance of state-of-the-art tracking methods on all benchmarks and sets a new state-of-the-art on the large-scale BDD100K MOT benchmark, while introducing negligible computational overhead to the detector.

Modelling De novo programming within Simon’s satisficing theory: Methods and application in designing an optimal offshore wind farm location system

De novo programming (DNP) is an efficient technique for optimal system design. This paper explores the ability to link the DNP technique with Simon’s satisficing theory to deal with a system design that is satisfactory rather than optimal. To achieve this aim, the ideal vector is replaced by an aspiration-level vector, and the solutions are determined by minimising the Lp-distance metric between the aspiration level and the feasible objective region. To generate a satisficing solution, we develop two models (weighted DNP (W-DNP) and Chebyshev DNP (C-DNP)) based on goal programming techniques. To achieve equilibrium between the solutions obtained from W-DNP and C-DNP, an extended DNP (E-DNP) model is proposed. Moreover, to deal with uncertainty and give decision makers more flexibility to incorporate their preferences, we consider the concept of penalty function (PF) with DNP and propose DNP type models with penalty functions (DNP-PFs). An illustrative example is adopted to show the usefulness of the proposed approach over the standard DNP. We also conduct a hypothetical application to Italian offshore wind farm locations to assess and validate the proposed formulations for solving real-world problems. To check the stability of the obtained results, the impact of the weights on the obtained solution is detected with a weight–space analysis. The results confirm the proposed methodologies and show that they can assist decision makers in determining the optimal location under uncertain aspiration levels.

Transformer-Based Visual Object Tracking with Global Feature Enhancement

With the rise of general models, transformers have been adopted in visual object tracking algorithms as feature fusion networks. In these trackers, self-attention is used for global feature enhancement. Cross-attention is applied to fuse the features of the template and the search regions to capture the global information of the object. However, studies have found that the feature information fused by cross-attention does not pay enough attention to the object region. In order to enhance cross-attention for the object region, an enhanced cross-attention (ECA) module is proposed for global feature enhancement. By calculating the average attention score for each position in the fused feature sequence and assigning higher weights to the positions with higher attention scores, the proposed ECA module can improve the feature information in the object region and further enhance the matching accuracy. In addition, to reduce the computational complexity of self-attention, orthogonal random features are introduced to implement a fast attention operation. This decomposes the attention matrix into the product of a random non-linear function between the original query and key. This module can reduce the spatial complexity and improve the inference speed by avoiding the explicit construction of a quadratic attention matrix. Finally, a tracking method named GFETrack is proposed, which comprises a Siamese backbone network and an enhanced attention mechanism. Experimental results show that the proposed GFETrack achieves competitive results on four challenging datasets.

PECULIARITIES OF CADASTRAL WORKS IN RESPECT OF LAND PLOTS OCCUPIED BY WATERLOGGED QUARRIES

The article focuses on the characterization of waterlogged quarries: their emergence, functioning and legal aspects of use; the issues of state registration of land plots occupied by waterlogged quarries are considered. The article substantiates the composition of cadastral works required for registration of a real estate object with the State cadastral register, and reviews the methods and regulatory requirements for the accuracy of determining the boundaries of land plots. It is recommended to carry out cadastral works taking into account and controlling the ecological and sanitary-epidemiological condition of a waterlogged quarry, which will provide the necessary information to public authorities, local self-government bodies, individuals and legal entities in the course of regulation and management of land relations for the organization of rational use and protection of land plots occupied by waterlogged quarries.

Tight Focusing of Circular Partially Coherent Radially Polarized Circular Airy Vortex Beam

The tight focusing properties of circular partially coherent radially polarized circular Airy vortex beams (CPCRPCAVBs) are theoretically studied in this paper. After deriving the cross-spectral density matrix of CPCRPCAVBs in the focal region of a high-NA objective, numerical calculations were performed to indicate the influence of the topological charge of the vortex phase on intensity distribution, degree of coherence and degree of polarization of the tightly focused beam. An intensity profile along the propagation axis shows that a super-length optical needle (~15 λ) can be obtained with a topological charge of 1, and a super-length dark channel (~15 λ) is observed with a topological charge of 2 or 3. In the focal plane, the rise in the number of topological charge does not distort the shapes of the coherence distribution pattern and the polarization distribution pattern, but enlarges their sizes.

Few-shot classification with multisemantic information fusion network

Metric-based methods aim to predict class labels by computing the similarity between samples using distance functions, which is the mainstream approach to few-shot learning. However, the limited representational space of feature vectors and appearance variations among congenetic samples still present challenges. We propose a Multisemantic Information Fusion Network (MIFN) to address these problems. A Lower-level Feature generator (LF-generator), which is an unsupervised module, adaptively activates high-response regions of objects to introduce discriminative semantic details. Meanwhile, a Higher-level Feature extractor (HF-extractor) learns global semantic information with human cognition to minimise the impact of appearance variations. We integrate the coarse outputs of these two modules, which complement each other to jointly promote more precise predictions. Furthermore, considering the importance of prototypes, we redefine the sampling strategy of the triplet loss and utilise it as an auxiliary loss to sharpen the decision boundary at the prototype level, facilitating subsequent classification. Our experimental results demonstrate the competitiveness of our approach in both general few-shot classification (mini-ImageNet and tiered-ImageNet) and cross-domain problems (CUB, Caltech-101, Stanford-dogs, and Stanford-cars) with minimal bells and whistles.

Anisotropic Multitentacle Janus Particles Synthesized by Selective Asymmetric Growth.

Anisotropic colloidal particles with asymmetric morphology possess functionally rich heterogeneous structures, thus offering potential for intricate superstructures or nanodevices. However, it is a challenge to achieve controlled asymmetric surface partitioned growth. In this work, an innovative strategy is developed based on the selective adsorption and growth of emulsion droplets onto different regions of object which is controlled by wettability. It is found that the emulsion droplets can selectively adsorb on the hydrophilic surface but not the hydrophobic one, and further form asymmetric tentacle by the interfacial sol-gel process along its trajectory. Janus particles with an anisotropic shape and multitentacle structure are achieved via integration of emulsion droplet (soft) and seed (hard) templates. The size and number of tentacles exhibit tunability mediated by soft and hard templates, respectively. This general strategy can be expanded to a variety of planar substrates or curved particles, further confirming the correlation between tentacle growth and Brownian motion. Most interestingly, it can be employed to selectively modify one region of surface partitioned particles to achieve an ABC three-component Janus structure.

Artificial intelligence in skeletal metastasis imaging

In the field of metastatic skeletal oncology imaging, the role of artificial intelligence (AI) is becoming more prominent. Bone metastasis typically indicates the terminal stage of various malignant neoplasms. Once identified, it necessitates a comprehensive revision of the initial treatment regime, and palliative care is often the only resort. Given the gravity of the condition, the diagnosis of bone metastasis should be approached with utmost caution. AI techniques are being evaluated for their efficacy in a range of tasks within medical imaging, including object detection, disease classification, region segmentation, and prognosis prediction in medical imaging. These methods offer a standardized solution to the frequently subjective challenge of image interpretation.This subjectivity is most desirable in bone metastasis imaging. This review describes the basic imaging modalities of bone metastasis imaging, along with the recent developments and current applications of AI in the respective imaging studies. These concrete examples emphasize the importance of using computer-aided systems in the clinical setting. The review culminates with an examination of the current limitations and prospects of AI in the realm of bone metastasis imaging. To establish the credibility of AI in this domain, further research efforts are required to enhance the reproducibility and attain robust level of empirical support.

Efforts in ensuring safety and transparency in space activities in the Asia-Pacific

This paper introduces the efforts to ensure the safety and transparency of space activities in the Asia-Pacific region, mainly based on the study results of the National Space Legislation Initiative (NSLI). NSLI is an initiative under the Asia-Pacific Regional Space Agency Forum (APRSAF) for exchanging information and enhancing the capacity of space policy and law. NSLI's main activity is the joint study of national space legislation for the submission of a report to the United Nations Committee on the Peaceful Uses of Outer Space (UNCOPUOS). NSLI's study includes topics relating to safety and transparency in space activities, such as space debris mitigation measures and registration of space objects. This paper introduces the status of such efforts by nine NSLI countries, mainly based on the information in the first phase NSLI report submitted in 2021 and discusses the way forward.

Memory Network With Pixel-Level Spatio-Temporal Learning for Visual Object Tracking

Making full use of temporal and spatial information is critical to cope with the appearance changes of objects in visual object tracking. However, existing methods in the tracking field, which employ a memory network at frame level to learn this information, bring redundancy and cannot build long-term relationships among historical frames due to the limited memory size. In this paper, we propose a novel memory network, Pixel-level Spatio-Temporal Memory (PSTM), which organizes object features in an efficient way to leverage temporal and spatial context information. Specifically, PSTM is constructed and updated by a memory writer, which includes a pixel-level updating strategy to maintain the temporal consistency and dynamically memorize the noteworthy variations. Furthermore, in order to exploit relationships between the object and search region and precisely estimate the state of the object, we propose a memory reader, Pixel-wise Matching and Refinement module (PMR), and model spatial context without a complex manual-designed mechanism. Comprehensive experiments and comparisons on challenging large-scale benchmarks, including GOT-10k, TrackingNet, LaSOT, OTB2015, VOT2020, and NfS, have demonstrated the effectiveness of our proposed method, which performs favorably against state-of-the-art trackers.

DANet: Dual-Branch Activation Network for Small Object Instance Segmentation of Ship Images

In maritime scenes, instance segmentation of small object ships is of vital importance. Small ship objects in images have the characteristics of smaller size, lower image cover rate and fewer appearance features. However, existing instance segmentation methods fail to recognize and segment them and can cause missed ship segmentation. To this end, we propose a dual-branch activation network (DANet) for small object instance segmentation of ship images. DANet consists of a Feature Encoding, a Dual Mask Branch, and a Dual Activation Branch. The Feature Encoding adopts feature refinement and a pyramid structure to obtain more fine-grained features. The proposed Dual Mask Branch extracts dual-path mask features for encoding small object information. We propose a Dual Activation Branch to activate more small object regions and generate instance features. Furthermore, we build the Small ShipInsSeg dataset from a total of 5,256 images and 11,612 instances. The experiments show that DANet outperforms the SparseNet baseline and achieves state-of-the-art performance. Additionally, our method achieves a good trade-off between accuracy and speed.

Module of Axis-based Nexus Attention for weakly supervised object localization

Weakly supervised object localization tasks remain challenging to identify and segment an entire object rather than only discriminative parts of the object. To tackle this problem, corruption-based approaches have been devised, which involve the training of non-discriminative regions by corrupting (e.g., erasing) the input images or intermediate feature maps. However, this approach requires an additional hyperparameter, the corrupting threshold, to determine the degree of corruption and can unfavorably disrupt training. It also tends to localize object regions coarsely. In this paper, we propose a novel approach, Module of Axis-based Nexus Attention (MoANA), which helps to adaptively activate less discriminative regions along with the class-discriminative regions without an additional hyperparameter, and elaborately localizes an entire object. Specifically, MoANA consists of three mechanisms (1) triple-view attentions representation, (2) attentions expansion, and (3) features calibration mechanism. Unlike other attention-based methods that train a coarse attention map with the same values across elements in feature maps, MoANA trains fine-grained values in an attention map by assigning different attention values to each element. We validated MoANA by comparing it with various methods. We also analyzed the effect of each component in MoANA and visualized attention maps to provide insights into the calibration.

Self-training based augmented reality for robust 3D object registration and task assistance

Typical augmented reality (AR) technologies superimpose virtual information onto the real environment for task guidance and assistance by tracking fiducial markers. However, these have narrow recognition ranges due to the markers, and it is difficult to track markers when occlusion occurs. Therefore, it is challenging to seamlessly register real objects in AR scenes captured by the camera without requiring fiducial markers. Deep learning–based approaches to 3D object registration and task assistance have recently indicated promising directions for superimposing 3D virtual objects onto corresponding real objects or their surroundings seamlessly. However, since supervised learning-based object registration requires many pose annotations on images for the locations and orientations of real objects, this process is a time-consuming and labor-intensive task. Thus, this paper proposes a new self-training based augmented reality method for seamless real object registration and task assistance without time-consuming pose annotations. As the industrial components or objects have their 3D geometric models, the proposed approach utilizes them as their synthetic data. The automatically generated synthetic dataset is used for training the deep learning model, which outputs a colored coordinate map of objects to be detected. 2D–3D key point matching between the 2D colored coordinate map and 3D colored coordinate model is then performed for 3D object registration in an AR environment. In particular, self-training is performed for fine-tuning the deep learning model by generating pseudo-labels with the highest probability for mask intersection over union (IoU) in the unlabeled real object dataset. We compared the proposed approach with previous studies to confirm its effectiveness and advantages. Also, we performed ablation studies by training a custom dataset based on various industrial 3D virtual models. Finally, we present AR-based task guidance and assistance, such as industrial robot manipulation.

A deep model towards accurate boundary location and strong generalization for medical image segmentation

Accurate medical image segmentation plays a crucial role in computer-assisted diagnosis and monitoring. However, due to the complexity of medical images and the limitations of image acquisition, most of the current segmentation models still face the challenges of imprecise boundary positioning and poor generalization ability, which greatly affect their application in clinical practices. To address these issues, we propose a new deep learning-based medical image semantic segmentation network (BLG-Net). To improve the accuracy of object boundary location, a Local–global module (Logo) is proposed. It uses the combination of a self-attention mechanism and a cross-attention mechanism to model intra-sample and inter-sample long-range dependencies from local and global perspectives, respectively, to help recover the spatial structures. To further enhance the generalization ability of the BLG-Net, we develop a generalization enhancement module (GEM), which can adaptively adjust the classifier weights of the network according to the features of the test images extracted during the test process to alleviate the distribution shift issues. Besides, we design a multi-granularity loss (MGL) function, which can guide the network to learn the object region and its boundary details at the pixel, patch, and map levels, respectively. Extensive experiments on five medical image segmentation tasks demonstrate the improvements of our method. We further verify the model’s generalization ability to unseen data on two other popular datasets with different modalities.

ExistenceMap-PointPillars: A Multifusion Network for Robust 3D Object Detection with Object Existence Probability Map.

Recognition of surrounding objects is crucial for ensuring the safety of automated driving systems. In the realm of 3D object recognition through deep learning, several methods incorporate the fusion of Light Detection and Ranging (LiDAR) and camera data. The effectiveness of the LiDAR-camera fusion approach is widely acknowledged due to its ability to provide a richer source of information for object detection compared to methods that rely solely on individual sensors. Within the framework of the LiDAR-camera multistage fusion method, challenges arise in maintaining stable object recognition, especially under adverse conditions where object detection in camera images becomes challenging, such as during night-time or in rainy weather. In this research paper, we introduce "ExistenceMap-PointPillars", a novel and effective approach for 3D object detection that leverages information from multiple sensors. This approach involves a straightforward modification of the LiDAR-based 3D object detection network. The core concept of ExistenceMap-PointPillars revolves around the integration of pseudo 2D maps, which depict the estimated object existence regions derived from the fused sensor data in a probabilistic manner. These maps are then incorporated into a pseudo image generated from a 3D point cloud. Our experimental results, based on our proprietary dataset, demonstrate the substantial improvements achieved by ExistenceMap-PointPillars. Specifically, it enhances the mean Average Precision (mAP) by a noteworthy +4.19% compared to the conventional PointPillars method. Additionally, we conducted an evaluation of the network's response using Grad-CAM in conjunction with ExistenceMap-PointPillars, which exhibited a heightened focus on the existence regions of objects within the pseudo 2D map. This focus resulted in a reduction in the number of false positives. In summary, our research presents ExistenceMap-PointPillars as a valuable advancement in the field of 3D object detection, offering improved performance and robustness, especially in challenging environmental conditions.

Siamese Trackers Based on Deep Features for Visual Tracking

Visual object tracking poses challenges due to deformation of target object appearance, fast motion, brightness change, blocking due to obstacles, etc. In this paper, a Siamese network that is configured using a convolutional neural network is proposed to improve tracking accuracy and robustness. Object tracking accuracy is dependent on features that can well represent objects. Thus, we designed a convolutional neural network structure that can preserve feature information that is produced in the previous layer to extract spatial and semantic information. Features are extracted from the target object and search area using a Siamese network, and the extracted feature map is input into the region proposal network, where fast Fourier-transform convolution is applied. The feature map produces a probability score for the presence of an object region and an object in a region, where the similarities are high to search the target. The network was trained with a video dataset called ImageNet Large Scale Visual Recognition Challenge. In the experiment, quantitative and qualitative evaluations were conducted using the object-tracking benchmark dataset. The evaluation results indicated competitive results for some video attributes through various experiments. By conducting experiments, the proposed method achieved competitive results for some video attributes, with a success metric of 0.632 and a precision metric of 0.856 as quantitative values.

Optimal allocation of agricultural water and land resources integrated with virtual water trade: A perspective on spatial virtual water coordination

Agricultural production consumes the majority of global freshwater resources. The worsening water scarcity has imposed significant stress on agricultural production when regions seek food self-sufficiency. To seek optimal allocation of spatial agricultural water and land resources in each water function zone of the objective region, a multi-objective optimization model was developed to tackle the trade-offs between the water-saving objective and the economic benefit objective considering virtual water trade (VWT). The cultivated area of each crop in each water function zone was taken into account as the decision variable, while a set of strong constraints were used to restrict land resources and water availability. Then, a decomposition-simplex method aggregation algorithm (DSMA) was proposed to solve this nonlinear, bounding-constrained, and multi-objective optimization model. Based on the quantitative analysis of the spatial blue and green virtual water in each agricultural product, the proposed methodology was applied to a real-world, provincial-scale region in China (i.e., Jiangsu Province). The optimized results provided 18 Pareto solutions to reallocate the land resources in the 21 IV-level water function zones of Jiangsu Province, considering four major rainy-season crops and two dry-season crops. Compared to the actual scenario, the superior scheme increased by 7.95% (5.6 × 109 RMB) for economic trade and decreased by 1.77% (2.0 × 109 m3) for agricultural water consumption. It was mainly because the potential of spatial blue and green virtual water in Jiangsu was fully exploited by improving spatial land resource allocation. The food security of Jiangsu could be guaranteed by achieving self-sufficiency in the superior scheme, and the total VWT in the optimal scheme was 2.2 times more than the actual scenario. The results provided a systematic decision-support methodology from the perspective of spatial virtual water coordination, yet, the methodology is widely applicable.

Phrase-Based Affordance Detection via Cyclic Bilateral Interaction

Affordance detection, which refers to perceiving objects with potential action possibilities in images, is a challenging task since the possible affordance depends on the person&#x0027;s purpose in real-world application scenarios. The existing works mainly extract the inherent human-object dependencies from image/video to accommodate affordance properties that change dynamically. In this paper, we explore to perceive affordances from a vision-language perspective, and consider the challenging phrase-based affordance detection task, i.e., given a set of phrases describing the potential actions, all the object regions in a scene with the same affordance should be detected. To this end, we propose a <b>C</b>yclic <b>B</b>ilateral <b>C</b>onsistency <b>E</b>nhancement Network (CBCE-Net) to align language and vision features in a progressive manner. Specifically, the presented CBCE-Net consists of a mutual guided vision-language module that updates the common features of vision and language in a progressive manner, and a <b>C</b>yclic <b>I</b>nteraction <b>M</b>odule (CIM) that facilitates the perception of possible interaction with objects in a cyclic manner. In addition, we extend the public Purpose-driven Affordance Dataset (PAD) by annotating affordance categories with short phrases. The extensive contrastive experimental results demonstrate the superior performance of our method over nine typical methods from four relevant fields in terms of both objective metrics and visual quality. The related code and dataset will be released at <uri>https://github.com/lulsheng/CBCE-Net</uri>. <i>Impact Statement</i>&#x2014;Affordance learning is concerned with the possible set of actions that an environment can offer to an actor. At present, almost all work focus on utilizing cues of human-object interactions in visual media such as images and videos to learn affordances. Few work explores affordance using language information, which is crucial to building more intelligent agents because we live a multimodal world. In this paper, we propose a phrase-based affordance detection task combining vision and language modalities. We first build a vision-language dataset based on the public Purpose-driven Affordance Dataset (PAD). Then we propose a method to fuse and align multi-modal features. Our method outperforms relevant methods from several other fields in terms of both objective metrics and visual quality. The proposed dataset and method can facilitate building more intelligent agents that can better comprehend the surrounding environment. For instance, the phase-based affordance detection can be used in building more intelligent home service systems. The agents can understand instructions of instructions without specific entities.