Target Specific Features Research Articles

Selective Information Flow for Transformer Tracking

Fully transformer-based one-stream trackers have demonstrated outstanding performance on challenging benchmark datasets over the past three years. These trackers enable bidirectional information flow between all target template and search region tokens to compute attention features without assessing its impact on the tracker’s discriminative capability. Our study found that the information flow from a large portion of background tokens in the search region diminishes the importance of the target-specific features of the template. Moreover, previous transformer-based trackers fail to consider cues from the dynamic background region, even though it contains information about distracting similar objects. To address the identified limitations in transformer tracking, we propose a novel Selective Information Flow Tracking (SIFTrack) framework to enhance the tracker’s discriminative capability by selectively allowing information flow between different types of token groups. In the early encoder layers of the proposed SIFTrack, interactions from all search tokens to target template tokens are blocked to enrich target-specific feature extraction. In the deeper encoder layers, search tokens are partitioned into target and non-target tokens based on their attention scores. Then, bidirectional flow from target search tokens to template tokens is allowed to capture the appearance changes of the target. In addition, by including tokens from the dynamic background, SIFTrack effectively avoids distractor objects by capturing cues from the area surrounding the target. The proposed SIFTrack demonstrated outstanding performance in challenging benchmarks, particularly excelling in the one-shot tracking benchmark GOT-10k, achieving an average overlap of 74.6%. The code, models, and results of this work are available at https://github.com/JananiKugaa/SIFTrack.

JFDI: Joint Feature Differentiation and Interaction for domain adaptive object detection

In unsupervised domain adaptive object detection, learning target-specific features is pivotal in enhancing detector performance. However, previous methods mostly concentrated on aligning domain-invariant features across domains and neglected integrating the specific features. To tackle this issue, we introduce a novel feature learning method called Joint Feature Differentiation and Interaction (JFDI), which significantly boosts the adaptability of the object detector. We construct a dual-path architecture based on we proposed feature differentiate modules: One path, guided by the source domain data, utilizes multiple discriminators to confuse and align domain-invariant features. The other path, specifically tailored to the target domain, learns its distinctive characteristics based on pseudo-labeled target data. Subsequently, we implement an interactive enhanced mechanism between these paths to ensure stable learning of features and mitigate interference from pseudo-label noise during the iterative optimization. Additionally, we devise a hierarchical pseudo-label fusion module that consolidates more comprehensive and reliable results. In addition, we analyze the generalization error bound of JFDI, which provides a theoretical basis for the effectiveness of JFDI. Extensive empirical evaluations across diverse benchmark scenarios demonstrate that our method is advanced and efficient.

Online discrimination-correction subnet for background suppression and dynamic template updating in Siamese visual tracking

The offline Siamese tracking methods have recently attracted much attention due to their impressive performance on speed and accuracy. However, they ignore the target and background information during tracking, thus lacking sufficient discrimination ability to handle complex background interference, and cannot robustly adapt to sharp appearance variations in challenging situations, such as scale variation, rotation, occlusion, and background clutter. To address the drawbacks above, we propose an online discrimination-correction subnet, which discriminates and corrects the tracking results obtained by the Siamese tracker through online learning, and further guides the template update. Firstly, a feature refinement module is designed to enhance the discriminative ability of features, which compresses and combines target features based on the importance of feature channels to obtain target-specific features. On this basis, an online discriminator module is proposed to learn the decision boundary between target and background through samples collected online to suppress background interference. In addition, a sample equilibrium strategy is proposed to alleviate the imbalance in the distribution of positive and negative samples during online training. Finally, to handle sharp target appearance variations, a robust template updating strategy is proposed to screen high-quality historical samples to generate dynamic templates. To evaluate the generalization ability and effectiveness of our method, we apply it to four classical Siamese trackers, namely SiamFC, SiamRPN++, SiamCAR, and TransT. Extensive experiments on seven challenging benchmarks demonstrate that our method significantly improves the baseline methods and achieves outstanding performance compared with state-of-the-art trackers.

Automatic purification of skylight spectrum based on an autoencoder

Abstract In the realm of astronomical spectroscopic observation, the purity of skylight spectra is crucial for accurate analysis, often complicated by interference from neighboring celestial objects. Addressing this challenge, a novel automatic purification algorithm is proposed in this study, leveraging the power of auto-coding mechanisms. By harnessing the inherent data features and shared representations between skylight and target objects, this method effectively isolates skylight spectra from residual signals of neighboring objects. Central to this approach is the utilization of an encoder-decoder framework. The encoder dynamically learns the combined features of skylight and target objects, generating respective coding vectors. These vectors are then utilized to subtract the target-specific features from the skylight coding space, facilitating the extraction of pure skylight characteristics. Subsequently, the decoder reconstructs the refined skylight data, effectively eliminating residual components associated with neighboring celestial bodies. A notable strength of this algorithm lies in its ability to perform feature extraction and purification autonomously, without relying on labeled data. Experimental validation conducted on LAMOST-DR5 datasets demonstrates the efficacy of the proposed method. By effectively removing spectral information from target objects within skylight spectra, the algorithm yields comparatively pure skylight spectra, paving the way for enhanced astronomical spectroscopic analyses.

A Multi-Objective Method Based on Tag Eigenvalues Is Used to Predict the Supply Chain for Online Retailers

E-commerce has grown quickly in recent years thanks to advancements in Internet and information technologies. For the majority of consumers, online shopping has emerged as a primary mode of shopping. However, it has become more challenging for businesses to satisfy consumer demand due to their increasingly individualized wants. To address the need for customized products with numerous kinds and small quantities, businesses must rebuild their supply chain systems to increase their efficiency and adaptability. The SI-LSF technique, which employs boosting learning in the target-relative feature space to lower the prediction error and enhance the algorithm's capacity to handle input-output interactions, is validated in this study using a genuine industrial dataset. The study successfully identifies the relationship between sales and sales as well as target-specific features by applying the multi-objective regression integration algorithm based on label-specific features to a real-world supply chain demand scenario.

PODB: A learning-based polarimetric object detection benchmark for road scenes in adverse weather conditions

Due to its insensitivity to light intensity and the capability to capture multidimensional information, polarimetric imaging technology has been proven to have advantages over traditional intensity-based imaging techniques for object detection tasks in adverse environmental conditions, particularly in road traffic scenarios. Recently, with the rapid development of artificial intelligence technology, deep learning (DL)-powered object detection techniques can further enhance recognition accuracy and algorithm robustness. This improvement is made possible by the ability of DL technology to leverage large datasets and extract deeper levels of target-specific features. However, constructing large-scale polarimetric datasets poses challenges as obtaining polarimetric information requires multiple captures of intensity images. In other words, the workload is several times higher compared to traditional imaging techniques. To address the current scarcity of polarimetric datasets and evaluate the practical performance of various algorithms on polarimetric datasets, this paper proposes a Polarimetric Object Detection Benchmark (PODB) dataset. The PODB provides an integrated quality evaluation framework for DL-based object detection algorithms in complex road scenes by incorporating polarimetric imaging. Besides, we conducted extensive object detection experiments using the PODB, which allowed for a comprehensive validation and performance evaluation of effective benchmark algorithms. Furthermore, a physics-based multi-scale image fusion cascaded object detection neural network model is proposed. By combining the multidimensional information provided by polarized images with an adaptive learning multi-decision object detection neural network model, the recognition accuracy of complex road scenes in adverse weather conditions has been improved by approximately 10%. Additionally, we anticipate that PODB will serve as an effective platform for evaluating and comparing the performance of object detection algorithms, as well as providing researchers with a baseline for future studies in developing new DL-based methods.

Discovery and learning of feature geometry in underwater acoustic sensing: Case studies and challenges in active sonar applications

Interpretable artificial intelligence (AI) and related machine learning (ML) techniques are gaining popularity for underwater acoustic sensing applications. However, interpretability of machine learnt features poses a fundamental challenge to successful application of these powerful data science techniques to underwater acoustics. Autonomous sonar target recognition is especially interesting from the data science perspective as beyond target-specific features, a sonar ping response typically includes significant interference from the environment, e.g., clutter, multipath scattering, etc. Such environmental interference, resulting from complex, dynamic, unpredictable, and often unknown factors, manifest as identifiable structures in acoustic color or alternate multi-dimensional feature representations that can lead to machine learning and classification errors. In this talk, we will discuss these challenges commonly encountered in underwater acoustics using case studies from field experiments in active sonar target recognition. Some results will also include physics-driven simulations in this domain to provide robust ground truths. In particular, we will posit how braided feature geometry and its representation, as well geometry of overlap between features can render a sonar target feature informative, explainable, and discoverable. [Work funded by the ONR Grant Nos. N000142112420 and N000142312503 and DoD Navy (NEEC) Grant No. N001742010016.]

Commonsense-based adversarial learning framework for zero-shot stance detection

Zero-shot stance detection (ZSSD) aims to identify the stance for a diverse range of topics with limited or no training data. It is more suitable for evolving real-world topics that are constantly evolving. However, ZSSD faces major challenges which are the lack of unseen target information and the inability to generalize training information to unseen targets effectively. To overcome these challenges, we propose a commonsense-based adversarial learning framework that comprises a commonsense graph encoder and a feature separation adversarial network. Specifically, our approach utilizes an external commonsense graph encoder to learn unseen target information. Moreover, we design a novel feature separation adversarial network to learn target-invariant and target-specific features, which enhances the model’s ability to reason beyond the seen targets. Experiments on two benchmark datasets show that our proposed framework achieves state-of-the-art performance.

Analysis of braid manifolds using uniform manifold approximation and projection

Autonomous sonar target recognition has been an ongoing challenge in the underwater signal processing community for many decades. Aspect-dependent target scattering and waveguide propagation causes target features to overlap, which increases the difficulty of constructing feature dictionaries. Current machine learning algorithms can suffer from an inability to track morphing target-specific features. Other ”black-box” machine learning algorithms produce results that are not explainable. We seek to extend previous work on creating feature representations using braid manifolds. Specifically, we use Uniform Manifold Approximation and Projection (UMAP) as a dimensionality reduction technique to search for underlying features that lay on a manifold. UMAP is an unsupervised learning algorithm that is built upon Riemannian geometry and fuzzy simplicial sets. The low dimensional representation embedding that UMAP produces is computed using both stochastic approximate nearest neighbor search and stochastic gradient descent with negative sampling. The performance of UMAP embeddings will be compared against other common dimension reduction algorithms and evaluated using various classifiers. [The authors would like to thank the Office of Naval Research for funding under Grant No. N00014-21-1-2420 and the University of Iowa.]

NullSpaceRDAR: Regularized discriminative adaptive nullspace for object tracking

Recently, discriminative-based and Siamese-based trackers have achieved outstanding performance on most tracking benchmarks. However, these trackers use the pre-trained backbone networks that have been mainly designed for classification to extract target-specific features without taking into consideration the visual object tracking task. In this paper, we propose NullSpaceRDAR, a novel tracker that learns a robust target-specific feature representation specifically designed for object tracking. This feature representation is learned by projecting the traditional backbone feature space onto a novel discriminative nullspace that is used to regularize the backbone loss function. We refer to the discriminative nullspace herein as joint-nullspace. The same target features (i.e., target-specific) in the proposed joint-nullspace are collapsed into a single point, and different target-specific features are collapsed into different points. Consequently, the joint-nullspace forces the network to be sensitive to the object’s variations from the same class (i.e., intra-class variations). Moreover, a modified adaptive loss function is developed for bounding box estimation to select the most suitable loss function from a super set family of loss functions based on the training data. This makes NullSpaceRDAR more robust to different challenges such as occlusions and background clutter. Extensive experiments have been conducted on six benchmarks to evaluate NullSpaceRDAR: OTB100, VOT variations (VOT2018, VOT2019, and VOT2020), LaSOT, TrackingNet, UAV123, and GOT10k. The results show that NullSpaceRDAR outperforms the state-of-the-art trackers.

Abstract 4062: Activity of 3 Bromopyruvate in human tumor primary culture 3-dimensional explants using ex vivo analysis of programmed cell death

Abstract Altered metabolism is a hallmark of human malignancy. The “Warburg Effect “reflecting increased glycolytic activity in the presence of oxygen represents an opportunity to target specific features of cancer to induce selective toxicity to malignantly transformed cells. 3-bromopyruvate (3-BP) is a pyruvate analogue that inhibits hexokinase II at the mitochondrial level resulting in cancer cell death. Numerous studies have confirmed activity for 3-BP in human SW480 and HT 29 colon carcinoma, glioblastoma (U373MG), HTP-1 leukemic and other cell lines. Our interest in exploring the activity of 3-BP in human tumor primary culture explants led to the application of the ex vivo analysis of programmed cell death in 19 patient tumors. The intent was to develop concertation ranges for future studies and explore disease-specific activity. Methods: Fresh Sterile tumor samples were mechanically and enzymatically disaggregated and 3-dimensional organoids of desired size isolated by density centrifugation. Blood samples were isolated by density centrifugation. Cell suspensions in modified RPMI media were exposed to serial dilutions of 3-BP for 72 hours and then assessed for cell death by morphology, staining characteristics and ATP content by luciferase. Viability was scored against saline-exposed controls, normalized to 100%. Results: 3BP reveals activity in human tumor explants. Activity varied by cancer type. Hematologic tumors including NHL, CLL and AML were found sensitive as were several solid tumors including GI and lung cancers. Of interest lower degrees of activity were observed in breast and ovarian cancers that are generally considered chemo-sensitive . Activity in 2 fibrolamellar cancers could suggest a novel therapy for this drug refractory neoplasm. Studies are underway examining 3-BP correlations with other classes of drugs and metabolic inhibitors as well as 3-BP combinatorial potential. Supported by the Fibro fighters Foundation Citation Format: Adam Jeremiah Nagourney, Paulo D'Amora, Steven Scott Evans, Paula Bernard, Federico Francisco, Young Ko, Thomas Stockwell, Maxine Johnson, Robert Alan Nagourney. Activity of 3 Bromopyruvate in human tumor primary culture 3-dimensional explants using ex vivo analysis of programmed cell death [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2022; 2022 Apr 8-13. Philadelphia (PA): AACR; Cancer Res 2022;82(12_Suppl):Abstract nr 4062.

Robust visual tracking algorithm with coattention guided Siamese network

In recent years, Siamese trackers locate the target based on the cross-correlation results of the template and the search branch. Siamese trackers emphasized better features embedding and similarity matching. The latest work focuses on enriching the feature representation of the template. However, it is still difficult for the tracker to eliminate the side effect of the distractors in the search region, resulting in the decline in tracking accuracy. We propose a coattention guided Siamese network (CGS) for visual tracking from two aspects. To mine the inherent correlation between the template and search branch, a coattention block is proposed that can guide the target-aware feature learning during the tracking process. To wisely choose the target-specific features, a self-gated mechanism is exploited on the template features, which can adaptively reallocate the confidence of channels and strengthen the feature representation. We conduct extensive experiments on OTB2015, VOT2016, and GOT-10K benchmarks to evaluate the performance of the proposed method. Experimental results demonstrate that our CGS network performs favorable against other state-of-the-art methods. Moreover, our CGS runs at 79FPS, far above the real-time requirement.

Defining the Critical DNA Features Targeted by RNA Polymerase I Core Factor

RNA Polymerase I(Pol I) is one of three essential DNA dependent RNA polymerases in eukaryotes and is responsible for synthesizing ribosomal RNA. A critical and essential Pol I transcription factor in yeast is Core Factor (CF) which binds to a ~24 bp region in the rDNA promoter called the Core Element (CE). CF plays fundamental roles in the Pol I transcription process helping to recruit Pol I and open Pol I promoter DNA before initiation. Previously, it was not yet known how CF precisely recognized the CE. When interacting with DNA, proteins use two main mechanisms: i) base‐pair readout, and ii) shape/structural readout. Base‐readout is the most common and is dominated by hydrogen bonding between the amino acid residues and base‐pair hydrogen bond donors and acceptors. The second mechanism centers around DNA shape readout where proteins target specific DNA features such as curvature, bendability, and groove width. Previous studies from our lab have shown that CF and its human orthologue, Selectivity Factor 1 (SL1), use an evolutionarily conserved mechanism to target DNA which is governed by interactions with the GC minor groove, a unique surface rarely targeted by DNA‐binding proteins. To further understand the extent of structural recognition as well as specific DNA features CF may be using to interact with CE, we have employed a variety of selection‐based methods to resolve the structural rules governing CF’s interaction with DNA. Our findings are consistent with a model that CF‐CE interaction is governed primarily by DNA shape‐based structural features rather than sequence.

Learning Enhanced Feature Responses for Visual Object Tracking.

Visual object tracking is an important topic in computer vision, which has successfully utilized pretrained convolutional neural networks, such as VGG and ResNet. However, the features extracted by these pretrained models are high dimensional, and the redundant feature channels reduce target localization and scale estimation precision, leading to tracking drifting. In this paper, a novel visual object tracking method, called learning enhanced feature responses tracking (LEFRT), is proposed, which adopts the target-specific features to enhance target localization and scale estimation responses. First, a channel attention module, called target-specific network (TSNet), is presented to reduce the redundant feature channels. Secondly, the scale estimation network (SCENet) is introduced to extract spatial structural features to generate a more precise response for the scale estimation. Extensive experiments on six tracking benchmarks, including LaSOT, GOT-10k, TrackingNet, OTB-2013, OTB-2015, and TC-128, demonstrate that the proposed algorithm can effectively improve the precision and speed of visual object tracking. LEFRT achieves 90.4% precision and a 71.2% success rate on the OTB-2015 dataset, improving the tracking methods based on the pretrained features.

Discriminative Siamese Complementary Tracker With Flexible Update

The offline generative Siamese trackers are equipped with the pre-defined anchors and the fixed target template. They overlook the target-background discriminative information, and lack the flexible target-specific update strategy. To overcome above drawbacks, we propose an adaptive and discriminative Siamese complementary tracking network with flexible update scheme. It consists of three collaborate subnetworks: anchor-free Siamese attention classification and regression subnetwork, online discriminative learning with multi-attention and multi-peak suppression, classifier guided template update subnetwork. All of them are interdependent and complementary to enhance each other for accurate target location. Specifically, an anchor-free multi-attention Siamese tracking subnetwork directly classifies the corresponding image patches with reliability assessment, and cascaded regresses the bounding boxes to progressively refine the predicting accuracy. Its evaluation is flexible and general with both proposal and anchor free in per-pixel prediction manner. Then, we integrate an online discriminative classifier optimizing module as a complementary subnetwork. It introduces spatial-temporal attention mechanism to fully explore multi-view multi-scale target-specific features, and evaluates multi-peak suppression to obtain a single centered peak response map. Its classified results can be fused with Siamese classification branch for accurate target location. Finally, the template update subnetwork is guided by the online discriminative classification scores. Extensive experiments on recent tracking datasets verify its top-ranked tracking accuracy and robustness against some state-of-the-art trackers.

Hybrid P4 Programmable Pipelines for 5G gNodeB and User Plane Functions

This paper focuses on hybrid pipeline designs for User Plane Function and next-generation NodeB leveraging target-specific features and an insightful discussion of P4 and target challenges and limitations. The entire or disaggregated UPF runs on P4 targets and allocates packet processing data paths in P4 hardware or DPDK/x86 software based on flow characteristics (e.g., heavy hitters) and QoS requirements (e.g., low-latency slices). For the hybrid gNodeB, most packet processing is executed in commodity Tofino hardware, while unsupported functions such as Automatic Repeat Request and cryptography are performed in DPDK/x86. We show that our hybrid UPF improves the scalability by 18× and reduces latency up to 50%. The results also suggest that careful traffic allocation to pipeline targets is required to optimize each target's strength and avoid processing delays. Finally, we demonstrate a QoS-oriented application of the hybrid UPF and present gNodeB buffer service benchmarks.

GaitSense: Towards Ubiquitous Gait-Based Human Identification with Wi-Fi

Gait, the walking manner of a person, has been perceived as a physical and behavioral trait for human identification. Compared with cameras and wearable sensors, Wi-Fi-based gait recognition is more attractive because Wi-Fi infrastructure is almost available everywhere and is able to sense passively without the requirement of on-body devices. However, existing Wi-Fi sensing approaches impose strong assumptions of fixed user walking trajectories, sufficient training data, and identification of already known users. In this article, we present GaitSense , a Wi-Fi-based human identification system, to overcome the above unrealistic assumptions. To deal with various walking trajectories and speeds, GaitSense first extracts target specific features that best characterize gait patterns and applies novel normalization algorithms to eliminate gait irrelevant perturbation in signals. On this basis, GaitSense reduces the training efforts in new deployment scenarios by transfer learning and data augmentation techniques. GaitSense also enables a distinct feature of illegal user identification by anomaly detection, making the system readily available for real-world deployment. Our implementation and evaluation with commodity Wi-Fi devices demonstrate a consistent identification accuracy across various deployment scenarios with little training samples, pushing the limit of gait recognition with Wi-Fi signals.

Transvaginal Ultrasound-Guided Fine-Needle Aspiration Biopsy of Pelvic Lesions.

To assess adequacy of transvaginal ultrasound-guided fine-needle aspiration biopsy (TVUS-FNAB) for pathologic diagnosis of pelvic masses performed using onsite cytopathology consultation. In this Institutional Review Board approved, Health Insurance Portability and Accountability Act (HIPAA) compliant study, radiology records were retrospectively queried to identify patients who underwent TVUS-FNAB of a pelvic mass over a 11-year duration. TVUS-FNAB adequacy was determined by correlating cytopathology results with transvaginal ultrasound-guided core-needle biopsy (TVUS-CNB) or surgical pathology results when available, and with clinical diagnostic confidence when additional pathology confirmation was not available. Secondary analysis included patient age, history of hysterectomy, or pelvic malignancy. Target-specific features analyzed included mass size, depth, location, and final pathologic diagnosis. Sixty patients underwent TVUS-FNAB of pelvic masses, 43 of which underwent FNAB only and 17 underwent both TVUS-FNAB and TVUS-CNB during the same procedure. TVUS-FNAB alone was adequate for diagnosis in 51 (85%) cases and addition of core-needle biopsy (CNB) achieved a diagnosis in additional 4 patients, increasing overall diagnostic accuracy to 92% (55/60). FNAB inadequacy had statistically significant association with increasing mass depth, occurrence of a minor intraprocedural complication, and decision to perform a CNB (P <.05). Number of FNAB passes, mass size, history of hysterectomy, and final diagnosis were not statistically significant predictors of FNAB adequacy. TVUS-FNAB has a high specimen adequacy rate when performed with an onsite cytopathologist and can be considered first-line approach for image-guided sampling of pelvic lesions with option to add CNB if preliminary cytopathologic review does not confirm sample adequacy.

Genome mining methods to discover bioactive natural products.

Covering: 2016 to 2021With genetic information available for hundreds of thousands of organisms in publicly accessible databases, scientists have an unprecedented opportunity to meticulously survey the diversity and inner workings of life. The natural product research community has harnessed this breadth of sequence information to mine microbes, plants, and animals for biosynthetic enzymes capable of producing bioactive compounds. Several orthogonal genome mining strategies have been developed in recent years to target specific chemical features or biological properties of bioactive molecules using biosynthetic, resistance, or transporter proteins. These “biosynthetic hooks” allow researchers to query for biosynthetic gene clusters with a high probability of encoding previously undiscovered, bioactive compounds. This review highlights recent case studies that feature orthogonal approaches that exploit genomic information to specifically discover bioactive natural products and their gene clusters.

Robust visual tracking via spatio-temporal adaptive and channel selective correlation filters

In recent years, Discriminative Correlation Filter (DCF) based tracking methods have achieved impressive performance in visual tracking. However, their excellent performance usually comes at the cost of sacrificing the computational speed. Furthermore, training correlation filters using high dimensional raw features may introduce the risk of severe over-fitting. To address the above issues, we propose Spatio-Temporal adaptive and Channel selective Correlation Filters (STCCF) for robust tracking. Specifically, we first select a set of target-specific features from high dimensional features via an effective channel selective scheme based on the Taylor expansion. Then, we reformulate the filter learning problem from ridge regression to elastic net regression to adaptively select the discriminative features inside the target bounding box at the spatial level. Moreover, we constrain the filters to be adaptive across temporal frames by learning a transformation matrix from the initial filters to the previous filters. In particular, with a specific spatio-temporal-channel constraint, STCCF can not only alleviate the over-fitting problem and reduce the computational cost, but also enhance the discriminability and interpretability of the learned filters. The proposed STCCF can be optimized by using a few iterations of Alternating Direction Method of Multipliers (ADMM). Experiments on six challenging datasets show that STCCF can achieve promising performance with fast running speed.