Position In Frame Research Articles

Fusion Key Frame Image Confidence Assessment of the Medical Service Robot Whole Scene Reconstruction

Facing the precise service and emergency rescue needs of medical service robots in irregular scene, in order to achieve better navigation and path planning for robots in service scenarios, for the whole reconstruction of the absolute scale service scenario, this article proposes a frame of whole scene three-dimensional (3D) point cloud reconstruction based on the fusion of scene depth estimation, confidence assessment, and pose tracking with monocular camera. The algorithm first collects the scene focus stack images under an initial viewing angle through the robot mobile terminal of camera. The absolute depth information of the scene is estimated on the server side, and the confidence level of the reconstructed image of the point cloud is evaluated, and non- uniform sampling is performed to reduce the influence of the error estimation. Based on the sparse key frame position information defined by monocular SLAM, the 3D reconstruction of the whole scene in absolute scale is realized through multi-perspective point cloud pose matching. It provides information of cloud reconstruction of scenic spots for target recognition and navigation of a medical service robot.

Context-Aware and Occlusion Handling Mechanism for Online Visual Object Tracking

Object tracking is still an intriguing task as the target undergoes significant appearance changes due to illumination, fast motion, occlusion and shape deformation. Background clutter and numerous other environmental factors are other major constraints which remain a riveting challenge to develop a robust and effective tracking algorithm. In the present study, an adaptive Spatio-temporal context (STC)-based algorithm for online tracking is proposed by combining the context-aware formulation, Kalman filter, and adaptive model learning rate. For the enhancement of seminal STC-based tracking performance, different contributions were made in the proposed study. Firstly, a context-aware formulation was incorporated in the STC framework to make it computationally less expensive while achieving better performance. Afterwards, accurate tracking was made by employing the Kalman filter when the target undergoes occlusion. Finally, an adaptive update scheme was incorporated in the model to make it more robust by coping with the changes of the environment. The state of an object in the tracking process depends on the maximum value of the response map between consecutive frames. Then, Kalman filter prediction can be updated as an object position in the next frame. The average difference between consecutive frames is used to update the target model adaptively. Experimental results on image sequences taken from Template Color (TC)-128, OTB2013, and OTB2015 datasets indicate that the proposed algorithm performs better than various algorithms, both qualitatively and quantitatively.

Adaptive Spatial Location With Balanced Loss for Video Captioning

Many pioneering approaches have verified the effectiveness of utilizing the global temporal and local object information for video understanding tasks and have achieved significant progress. However, existing methods utilize object detectors to extract all objects overall video frames. This may bring performance degradation due to the information redundancy both spatially and temporally. To address this problem, we propose an adaptive spatial location module for the video captioning task which dynamically predicts an important position of each video frame in the procedure of generating the description sentence. The proposed adaptive spatial location method not only makes our model focus on local object information, but also reduces time and memory consumption brought by the temporal redundancy in extensive video frames and improves the accuracy of generated description. Besides, we propose a balanced loss function to address the class imbalance problem existing in training data. The proposed balanced loss assigns different weight to each word of ground-truth sentence in the training process which can generate more diversified description sentences. Extensive experimental results on the MSVD and MSR-VTT dataset show that the proposed method achieves competitive performance compared to state-of-the-art methods.

Zero-marked valency alternations in Macedonian

Artykuł omawia alternacje walencyjne, nie posiadające jawnego wykładnika w języku macedońskim. Wydaje się, że jest to ważna cecha typologiczna, która odróżnia macedoński od innych języków słowiańskich. W pracy przedstawiono ogólny opis tendencji labilności w odniesieniu do cech składniowych, semantycznych i dystrybucyjnych czasownika. Wykryto i przeanalizowano około 150 czasowników, wchodzących w labilność zachowującą przy alternacji patiensa i agensa. Czasowniki typu pierwszego przy zmianie perspektywy zachowują miejsce dla patiensa, ale w innej pozycji syntaktycznej (Toj sedna – Tie go sednaa. ‘On siedział – Posadzili go’); typu drugiego zaś zachowują tematyczne miejsce dla agensa (Taa odi po poleto – Taa go odi poleto ‘Idzie przez pole – Idzie w pole’). Czasowniki zostały poklasyfikowane zgodnie z kryteriami składniowymi i semantycznymi po to, aby wskazać zdarzenia, które dopuszczają tego typu alternacje walencyjne oraz dominujące korelacje semantyczne, zachodzące między parami czasowników.

Visual ship tracking via a hybrid kernelized correlation filter and anomaly cleansing framework

Ship tracking from maritime visual sensing data (namely maritime surveillance videos) provides various kinematic maritime traffic information, which significantly benefits remote maritime traffic controlling and management, off-site law enforcement, etc. But, it is difficult to extract distinct ship visual features when the target ship is sheltered by the neighboring ship in the maritime images (or the images are shot in low visibility condition). To address the difficulty, we proposed a hybrid ship tracking framework via the help of kernelized correlation filter (KCF) and anomaly cleaning models (including curve fitting method and Kalman filter). First, we employed the KCF model to obtain raw ship trajectories in consecutive maritime images. Second, our ship tracker is accurately initialized (to be specific, ground truth ship position in the first frame is employed to initialize the ship tracker). Third, the Kalman filter is introduced to suppress the trivial ship position oscillations in the raw ship trajectories. We verified the proposed framework performance on the four typical maritime scenarios. The experimental results indicate that the proposed ship tracker showed more accurate ship tracking results compared to other four popular ship trackers in terms of average root mean square error (RMSE), mean absolute deviation (MAD) and mean square error (MSE). The research findings can help maritime traffic participants obtain visual on-spot maritime kinematic information, and thus further enhance maritime traffic safety.

Validation of Omega Subunit of RNA Polymerase as a Functional Entity.

The bacterial RNA polymerase (RNAP) is a multi-subunit protein complex (α2ββ’ω σ) containing the smallest subunit, ω. Although identified early in RNAP research, its function remained ambiguous and shrouded with controversy for a considerable period. It was shown before that the protein has a structural role in maintaining the conformation of the largest subunit, β’, and its recruitment in the enzyme assembly. Despite evolutionary conservation of ω and its role in the assembly of RNAP, E. coli mutants lacking rpoZ (codes for ω) are viable due to the association of the global chaperone protein GroEL with RNAP. To get a better insight into the structure and functional role of ω during transcription, several dominant lethal mutants of ω were isolated. The mutants showed higher binding affinity compared to that of native ω to the α2ββ’ subassembly. We observed that the interaction between α2ββ’ and these lethal mutants is driven by mostly favorable enthalpy and a small but unfavorable negative entropy term. However, during the isolation of these mutants we isolated a silent mutant serendipitously, which showed a lethal phenotype. Silent mutant of a given protein is defined as a protein having the same sequence of amino acids as that of wild type but having mutation in the gene with alteration in base sequence from more frequent code to less frequent one due to codon degeneracy. Eventually, many silent mutants were generated to understand the role of rare codons at various positions in rpoZ. We observed that the dominant lethal mutants of ω having either point mutation or silent in nature are more structured in comparison to the native ω. However, the silent code’s position in the reading frame of rpoZ plays a role in the structural alteration of the translated protein. This structural alteration in ω makes it more rigid, which affects the plasticity of the interacting domain formed by ω and α2ββ’. Here, we attempted to describe how the conformational flexibility of the ω helps in maintaining the plasticity of the active site of RNA polymerase. The dominant lethal mutant of ω has a suppressor mapped near the catalytic center of the β’ subunit, and it is the same for both types of mutants.

Human motion tracking and positioning for augmented reality

AR (Augmented reality) is a research hotspot in the current computer application field. AR technology enhances people’s understanding and experience of the real environment by adding virtual objects to real scenes to integrate virtual objects with the real environment. Aiming at the weak processing power of intelligent terminals and the characteristics of limited hardware resources, this paper proposes a more effective human motion feature extraction and descriptor algorithm. The feature point detection and positioning method suitable for intelligent terminals is proposed in a targeted manner, which solves the problem of mismatching of similar structures. In addition, this paper proposes an AR-oriented recursive tracking algorithm for human motion. The positional relationship of the current frame is calculated from the position of the previous frame. A combination of ORB (Oriented fast and Rotated Brief) feature descriptors and KLT (Kanade-Lucas-Tomasi) algorithm is adopted. The ORB feature descriptor matched by the first frame image and the reference image is tracked by the KLT tracking algorithm, and the feature descriptor of the previous frame is tracked in the current frame, thereby eliminating the phenomenon of virtual object jitter. The experimental results show that the recursive tracking scheme has better performance in time and precision than the detection tracking scheme.

Self-supervised representation learning by predicting visual permutations

We propose a self-supervised learning method to uncover the spatial or temporal structure of visual data by identifying the position of a patch within an image or the position of a video frame over time, which is related to Jigsaw puzzle reassembly problem in previous works. A Jigsaw puzzle can be seen as a shuffled sequence, which is generated by shuffling image patches or video frames according to an unknown permutation. The task of predicting the visual permutations can be used to train a learning system to capture structural information which is important for semantic-level tasks, such as object recognition and action recognition. To this end, we propose a multi-task learning framework where a group of principal tasks aims to predict the index of each sample in the original sequence, and a group of auxiliary tasks aims to predict the spatial or temporal relation of adjacent samples in the shuffled sequence. Our scheme can handle the whole space of permutations and is fairly scalable, and it is also generic to solve many problems such as self-supervised representation learning, relative attributes, and learning to rank. Our method achieves state-of-the-art performance on the STL-10 benchmarks for unsupervised representation learning, and it is competitive with state-of-the-art performance on UCF-101 and HMDB-51 as a pretraining method for action recognition. In addition, we apply the proposed method on age comparison task to prove it is generic to solve ranking problems.

Human activity recognition using improved dynamic image

In action recognition, the dynamic image (DI) approach is recently proposed to code a video signal to a still image. Since DI descriptor is strongly dependent on first frames, it cannot extract dynamics that do not occur in the first frames or even long dynamics. On the other hand, most of the video frames are not informative for the task of action recognition. Therefore, the authors' intuition is that the process of representing a video using all frames is inefficient. Thus, in this study, they proposed to remove the existing redundancy inside the frames and extract some processed informative images based on the information theory which are called key frames. The proposed method is capable enough to extract sufficient frames regardless of the duration and the position of frames in the entire video. Motivated by this method and DI, they proposed a novel key frames dynamic image (KFDI) approach. Experimental results on popular UCF11, Olympic Sports, and J-HMDB datasets show the superiority of the proposed KFDI approach compared to the DI in capturing long dynamics of videos for action recognition. Their experiments show KFDI improves the accuracy 2–6%compared to DI.

Теплові характеристики вузлів примикання вікон до цегляних та залізобетонних стін цивільних будівель на території Кіровоградської області

Existing design standards require that the temperature of the inner surface of the enclosure in the zones of high thermal conductivity should not fall below the temperature of the dew point. This phenomenon can occur in a place of window framing to the wall of residential and public buildings, which were massively produced in the second half of the twentieth century. Condensation can be avoided by raising the surface temperature of the inner window jamb extensions. This can be done by adding an additional insulation or displacement of the block frame window inside the building. The aim of the work is to calculate the appropriate values of the displacement of the block frame window inside the wall of different types for the climatic conditions of the Kirovograd region. The framing of metal-plastic windows to the walls of various structures that were produced in the second half of the twentieth century are analyzed. The insulation is performed in accordance with DBN B.2.6-31-2016. The lowest design temperature of the coldest five-day period within the territory of the Kirovograd Region is -24 °С. As a safety margin, the dew point temperature for residential and public buildings for various purposes is taken to be +11 °C. Finite element modeling in the THERM program environment made it possible to construct the temperature fields of the framing of metal-plastic windows to the walls of various structures and to obtain the very temperature of the inner window jamb extensions at various positions of the window frame within the wall. For each of the nodes examined, the possibility of condensation on the inner surfaces of the walls was established taking into account the position of window frame within the wall. As a result of the studies, it was found that the nodes of the side framing of the windows to the brick and panel walls of civil buildings without additional facade insulation do not meet the requirements for thermal reliability according to the condensation criterion. Facade insulation of existing or new walls, made in accordance with the requirements of DBN B.2.6-31-2016 "Thermal insulation of buildings", in almost all cases guarantees the absence of condensation on the internal surfaces of window jambs.

Unsupervised Deep Representation Learning for Real-Time Tracking

The advancement of visual tracking has continuously been brought by deep learning models. Typically, supervised learning is employed to train these models with expensive labeled data. In order to reduce the workload of manual annotation and learn to track arbitrary objects, we propose an unsupervised learning method for visual tracking. The motivation of our unsupervised learning is that a robust tracker should be effective in bidirectional tracking. Specifically, the tracker is able to forward localize a target object in successive frames and backtrace to its initial position in the first frame. Based on such a motivation, in the training process, we measure the consistency between forward and backward trajectories to learn a robust tracker from scratch merely using unlabeled videos. We build our framework on a Siamese correlation filter network, and propose a multi-frame validation scheme and a cost-sensitive loss to facilitate unsupervised learning. Without bells and whistles, the proposed unsupervised tracker achieves the baseline accuracy of classic fully supervised trackers while achieving a real-time speed. Furthermore, our unsupervised framework exhibits a potential in leveraging more unlabeled or weakly labeled data to further improve the tracking accuracy.

How Plants Influence Resilience of Salt Marsh and Mangrove Wetlands to Sea-Level Rise

This review evaluates the importance of plants and associated biological processes in determining the vulnerability of coastal wetlands to sea-level rise. Coastal wetlands occur across a broad sedimentary continuum from minerogenic to biogenic, providing an opportunity to examine the relative importance of biological processes in wetland resilience to sea-level rise. We explore how plants influence sediment accretion, elevation capital (vertical position in the tidal frame), and compaction or erosion of deposited material. We focus on salt marsh and mangrove wetlands, which occupy a similar physiographic niche and display similar physical and biological controls on resilience to sea-level rise. In both habitats, plants stabilize emergent mudflats and help sustain the wetland position in the tidal frame relative to ocean height through both surface and subsurface process controls on soil elevation. Plants influence soil elevations by modifying (1) mineral sediment deposition and retention, (2) organic matter contributions to soil volume, and (3) resistance to compaction and erosion. Recognition of the importance of plants in coastal wetland resilience to sea-level rise is key to accurate predictions about the future fate of salt marshes and mangrove forests and for development of effective management and restoration plans.

Improved YOLO v3 network-based object detection for blind zones of heavy trucks

Object detection for blind zones is critical to ensuring the driving safety of heavy trucks. We propose a scheme to realize object detection in the blind zones of heavy trucks based on the improved you-only-look-once (YOLO) v3 network. First, according to the actual detection requirements, the targets are determined to establish a new data set of persons, cars, and fallen pedestrians, with a focus on small and medium objects. Subsequently, the network structure is optimized, and the features are enhanced by combining the shallow and deep convolution information of the Darknet platform. In this way, the feature propagation can be effectively enhanced, feature reuse can be promoted, and the network performance for small object detection can be improved. Furthermore, new anchors are obtained by clustering the data set using the K-means technique to improve the accuracy of the detection frame positioning. In the test stage, detection is performed using the trained model. The test results demonstrate that the proposed improved YOLO v3 network is superior to the original YOLO v3 model in terms of the blind zone detection and can satisfy the accuracy and real-time requirements with an accuracy of 94% and runtime of 13.792 ms / frame. Moreover, the mean average precision value for the improved model is 87.82%, which is 2.79% higher than that of the original YOLO v3 network.

T1 -FLAIR imaging during continuous head motion: Combining PROPELLER with an intelligent marker.

The purpose of this work is to describe a T1 -weighted fluid-attenuated inversion recovery (FLAIR) sequence that is able to produce sharp magnetic resonance images even if the subject is moving their head throughout the acquisition. The robustness to motion artifacts and retrospective motion correction capabilities of the PROPELLER (periodically rotated overlapping parallel lines with enhanced reconstruction) trajectory were combined with prospective motion correction. The prospective correction was done using an intelligent marker attached to the subject. This marker wirelessly synchronizes to the pulse sequence to measure the directionality and magnitude of the magnetic fields present in the MRI machine during a short navigator, thus enabling it to determine its position and orientation in the scanner coordinate frame. Three approaches to incorporating the marker-navigator into the PROPELLER sequence were evaluated. The specific absorption rate, and subsequent scan time, of the T1 -weighted FLAIR PROPELLER sequence, was reduced using a variable refocusing flip-angle scheme. Evaluations of motion correction performance were done with 4 volunteers and 3 types of head motion. During minimal out-of-plane movement, retrospective PROPELLER correction performed similarly to the prospective correction. However, the prospective clearly outperformed the retrospective correction when there was out-of-plane motion. Finally, the combination of retrospective and prospective correction produced the sharpest images even during large continuous motion. Prospective motion correction of a PROPELLER sequence makes it possible to handle continuous, large, and high-speed head motions with only minor reductions in image quality.

Deep Reinforcement Learning-Based Progressive Sequence Saliency Discovery Network for Mitosis Detection In Time-Lapse Phase-Contrast Microscopy Images.

Mitosis detection plays an important role in the analysis of cell status and behavior and is therefore widely utilized in many biological research and medical applications. In this article, we propose a deep reinforcement learning-based progressive sequence saliency discovery network (PSSD)for mitosis detection in time-lapse phase contrast microscopy images. By discovering the salient frames when cell state changes in the sequence, PSSD can more effectively model the mitosis process for mitosis detection. We formulate the discovery of salient frames as a Markov Decision Process (MDP)that progressively adjusts the selection positions of salient frames in the sequence, and further leverage deep reinforcement learning to learn the policy in the salient frame discovery process. The proposed method consists of two parts: 1)the saliency discovery module that selects the salient frames from the input cell image sequence by progressively adjusting the selection positions of salient frames; 2)the mitosis identification module that takes a sequence of salient frames and performs temporal information fusion for mitotic sequence classification. Since the policy network of the saliency discovery module is trained under the guidance of the mitosis identification module, PSSD can comprehensively explore the salient frames that are beneficial for mitosis detection. To our knowledge, this is the first work to implement deep reinforcement learning to the mitosis detection problem. In the experiment, we evaluate the proposed method on the largest mitosis detection dataset, C2C12-16. Experiment results show that compared with the state-of-the-arts, the proposed method can achieve significant improvement for both mitosis identification and temporal localization on C2C12-16.

Contemporary distribution of benthic foraminiferal assemblages in coastal wetlands of south-eastern Australia

Benthic foraminifera identified from the upper intertidal zone of coastal mangrove and saltmarsh wetlands of south-eastern Australia exhibit spatial zonation that varies with elevation and corresponds to position in the tidal frame. This study describes relationships between benthic foraminiferal assemblages and environmental variables, such as wetland elevation, pore-water salinity, substrate organic matter content and sediment grainsize. Agglutinated taxa dominated higher in the tidal frame, whereas calcareous taxa increased in both relative abundance and diversity lower in the tidal frame. Species composition varied within and between study sites, reflecting local differences in other environmental variables. Three major spatial clusters were identified in the upper intertidal zone from a dataset consisting of 111 contemporary samples: a high-to mid-elevation assemblage consisting of differing abundances of Entzia macrescens and Trochammina inflata; a mid-elevation assemblage comprised of Arenoparrella mexicana in association with Trochammina inflata and Quinqueloculina seminula; and mid-to low-elevation assemblage characterised by Cribroelphidium excavatum, Quinqueloculina seminula, Ammobaculites spp. and Miliammina fusca. Multivariate analysis correlated foraminiferal assemblages with environmental variables showing that foraminiferal distribution in the region is constrained primarily by wetland elevation, but also by pore-water salinity. Foraminiferal assemblages derived from cluster analysis correlated with vegetation zones, implying that both foraminifera and vegetation are influenced by elevation and salinity. This paper seeks to demonstrate that the consistency in spatial zonation of assemblages provides greater confidence that contemporary distribution of foraminiferal assemblages can assist with the interpretation of past depositional conditions and palaeo-environmental reconstructions, providing insights into the evolution of coastal wetlands of south-eastern Australia over the Holocene.

Simplified support structure design for multi-rotor wind turbine systems

Abstract. In this study different multi-rotor wind turbine systems (MRSs) are designed in such a way that the space frame, forming the connection between rotor nacelle assemblies (RNAs) and the tower, is modeled as an ideal truss work. To dimension the tube diameters and wall thicknesses, a simplified load case is used with an adjusted safety factor for loads. This simplified approach allows for fast computations of a large variety of different support structure designs. By variation of rotor number, space frame topology, space frame depth and positioning of yaw bearings, it is possible to gain an understanding of the optimal MRS design. As such, the simplified approach is a preliminary step helping to choose a good design parameter combination for a more detailed and comprehensive analysis.

Hierarchical support vector machine for facial micro-expression recognition

The sample category distribution of spontaneous facial micro-expression datasets is unbalanced, due to the experimental environment, collection equipment, and individualization of subjects, which brings great challenges to micro-expression recognition. Therefore, this paper introduces a micro-expression recognition model based on the Hierarchical Support Vector Machine (H-SVM) to reduce the interference of sample category distribution imbalance. First, we calculated the position of the apex frame in the micro-expression image sequence. To keep micro-expression frames balanced, we sparsely sample the images sequence according to the apex frame. Then, the Low-level Descriptors of the region of interest of the micro-expression image sequence and the High-level Descriptors of apex frame are extracted. Finally, the H-SVM model is used to classify the fusion features of different levels. The experimental results on SMIC, CAMSE2, SAMM, and their composite datasets show that our method can achieve superior performance in micro-expression recognition.

Trapping of Suspended Sediment by Submerged Aquatic Vegetation in a Tidal Freshwater Region: Field Observations and Long-Term Trends

Widespread invasion by non-native, submerged aquatic vegetation (SAV) may modify the sediment budget of an estuary, reducing the availability of inorganic sediment required by marshes to maintain their position in the tidal frame. The instantaneous trapping rate of suspended sediment in SAV patches in an estuary has not previously been quantified via field observations. In this study, flows of water and suspended sediment through patches of invasive SAV were measured at three tidally forced, freshwater sites, all located within the Sacramento-San Joaquin Delta in California. An acoustic Doppler current profiler deployed from a roving vessel provided velocity and backscatter data used to quantify fluxes of both water and suspended sediment. Sediment trapping efficiency, defined as instantaneous net trapped flux divided by incident flux, was positive in 24 of 29 cases, averaging + 5%. Coupled with 3 years of measured sediment flux data at one site, this suggests that trapping averages 3.7 kg m−2 year−1. This estimate compares favorably with the mean mass accumulation rate of 3.8 kg m−2 year−1 estimated from dated sediment cores collected at the study sites. Long-term measurements made upstream reveal a strong negative trend (− 1.8% year−1) in suspended sediment concentration, and intra-annual changes in both suspended sediment concentration and percent fines. The large footprint and high spatial density of invasive SAV coupled with declining sediment supply are diminishing downstream suspended sediment concentrations, potentially reducing the resiliency of marshes in the Delta and lower estuary to future sea-level rise.

Target-centric angle variables in pursuit tracking

A method for constructing a dynamic local reference frame centered at the location of a pursuit target is proposed. It uses original fixed reference frame coordinates for three positions: the location of a moving target, the location of a pursuing entity, and a third reference point that defines the instantaneous target trajectory direction. The method involves rigid body geometric operations and creates a consistent local reference frame configuration having its origin at the target location. The local frame also has the target trajectory pointing in the positive x axis direction, and a defined azimuth plane descriptive of the current pursuer location. This plane is determined by a concluding local frame rotation around the set x axis, which specifies a particular y axis direction. If the pursuer position is not on the local frame’s azimuth plane, its perpendicular projection on to that plane serves as an additional position for use in defining angle variables. Three angle variables relevant to the pursuit dynamics (not all mutually independent) are obtained from the three coordinate values of the pursuer position in the local frame: an elevation angle, an azimuth angle, and a bearing angle that basically describes how far off course the pursuer position is from being directly behind the target’s current trajectory path position. To illustrate the method of obtaining these angles and to demonstrate their potential usefulness in characterizing pursuit behavior, example experimental data from a motor control and coordination pursuit task performance are presented and analyzed.