Object Identification Research Articles

Multi-model Deep Learning Approaches for Vehicle Speed Estimation

Accurate vehicle speed estimates are critical for traffic monitoring and management. Traditional speed estimation approaches, such as radar guns or pixel displacement methods, are resource-intensive and frequently struggle in real-time or dynamic traffic situations. This paper describes a combination of YOLO for real-time object identification and Long Short-Term Memory (LSTM) networks for vehicle speed prediction that overcomes these constraints. The ability of YOLO to recognize cars with high precision, along with the capability of LSTM in processing sequential data, makes this approach extremely good for calculating speed from video data. Using the VS13 dataset of 400 single-vehicle videos, YOLOV5 detects vehicles and generates bounding box areas for LSTM to estimate speed. The performance across 13 car models averaged 5.79 km/hr RMSE. The accuracy and generalizability of the model can be improved in the future by adding outside variables like weather, road conditions, and driver behavior.

Object identity representation occurs early in the archerfish visual system

Archerfish hunt by shooting a jet of water at aerial targets, a behavior used to study their visual processing by presenting a set of images on a screen above the water tank and observing the behavioral response. Building on this unique behavior, it was recently shown that archerfish can be trained to distinguish between different object categories by generalizing from examples. Analysis of the archerfish’s behavior revealed that the fish visual system relies on a small set of visual features for categorization and is more sensitive to object contours than to textures. To understand the neural basis of this object recognition, we investigated the neural representation of features and objects in the archerfish optic tectum using recording of single cells. We found that, although the optic tectum is an early stage of visual processing, a small population of neurons in this region contains information about the object category. This contrasts with the primate visual system, where the representation of objects emerges only at later stages of visual processing. These results suggest that early-stage feature extraction and object categorization in archerfish might represent a form of specialized visual processing. This contributes to a broader understanding of visual processing across taxa.

Vision of objects happens faster and earlier for location than for identity.

Visual perception of objects requires the integration of separate independent stimulus features, such as object identity and location. We ask whether the location and the identity of an object are processed with different efficiency for being consciously recognized and reported. Participants viewed a target letter at one out of several locations that were terminated by pattern masks at all possible locations. Participants reported the location of the target and/or its letter identity. Report performance as a function of the target duration before the mask is enabled to estimate the speed of visual processing and the minimum duration for processing to start. Visual processing was faster and started earlier for spatial location than for object identity, even though the processing of the features was (stochastically) independent. Together, these findings reveal an intrinsic preference of the human visual system for the perceptual processing of space as opposed to visual features such as categorical identity.

Overview of instrumental methods used in the field of food analysis

The emergence of new technologies for food production and current trends in the use of alternative sources of raw materials require improved approaches to the analysis of the chemical composition of food products. In the course of planning work to carry out analytical research, a number of challenges arise: identification of the research objects themselves; selection of research methods and tools; ensuring the representativeness and validity of the results obtained. When choosing and justifying research methods, it is necessary to rely on a large number of factors, which include both confirmation of the actual composition of food products, including the determination of unintentionally present substances, and the reliability of the data obtained, taking into account the selected instrumental methods. The purpose of this work is to generalize and systematize the essence and characteristics of the main methods of food analysis and evaluate existing approaches to the justification and application of instrumental methods of analysis in relation to new types of food products. This paper reviews the literature on instrumental research methods used to obtain values in the most representative international databases on the composition of food products (FAO/INFOODS Food Composition Databases, USDA NDL, Fineli, Frida), as well as databases of the chemical composition of food products in Russia, Japan and Australia. To search for descriptions and features of the use of analytical equipment and analytical methods, electronic library systems Web of Science, Scopus, Elibrary, ResearchGate, Google Scholar, Microsoft Academic, Science Direct were used. This review highlights the role of various research methods: photometric and electrophoretic, titrimetric, extraction, chromatographic, spectroscopic, immunoenzymatic, as well as those based on the polymerase chain reaction and the use of nuclear magnetic resonance, lateral flow and electropheresis.

Examining the physical and psychological effects of combining multimodal feedback with continuous control in prosthetic hands

Myoelectric prosthetic hands are typically controlled to move between discrete positions and do not provide sensory feedback to the user. In this work, we present and evaluate a closed-loop, continuous myoelectric prosthetic hand controller, that can continuously control the position of multiple degrees of freedom of a prosthesis while rendering proprioceptive feedback to the user via a haptic feedback armband. Twenty-eight participants without and ten participants with upper limb difference (ULD) were recruited to holistically evaluate the physical and psychological effects of the controller via isolated control and sensory tasks, dexterity assessments, embodiment and task load questionnaires, and post-study interviews. The combination of proprioceptive feedback and continuous control enabled more accurate position and force modulation than without proprioceptive feedback, and restored blindfolded object identification ability to open-loop discrete controller levels. Dexterity assessment and embodiment questionnaire results revealed no significant physical performance or psychological embodiment differences between control types, with the exception of perceived sensation questions, which were significantly higher (p < 0.001) for closed-loop controllers. Key differences between participants with and without ULD were identified, including increasingly lower perceived body completeness and heterogeneity in frustration in participants with ULD, which can inform future development and rehabilitation.

Bioinspired Tactile Object Identification Leveraging Deep Learning and Soft Body Compliance

Tactile object identification is a fundamental human skill, underlying several core aspects of human intelligence. Humans display a range of remarkable haptic skills, enabled by the synergistic interactions of the somatosensory system with higher‐level cognitive processes. In contrast, robotics’ haptic sensing solutions have historically lacked the ability to achieve human‐level perceptive capabilities, lacking in both the sensory system and its cognitive digital counterpart. Herein, part of this challenge is addressed by leveraging the success of the fields of soft robotics and deep learning to show how a soft robotic hand, equipped with low‐resolution tactile sensing, can be used to accurately identify a diverse set of objects. In particular, ROSE‐Net, a neural network that leverages multiple grasps to enable accurate pose‐invariant object recognition, is developed. The multi‐grasp haptic discrimination solution can lead to a significant increase in performance. The versatility and adaptability of this approach are also tested in two scenarios: a learning transfer scenario and a fault tolerance scenario. Finally, the framework is tested in an online discrimination task, where this approach is shown to naturally require additional grasps for objects that are more challenging to identify using a single grasp and low spatial resolution tactile sensing.

Play it Again, Sam? Reference-Point Formation and Product Differentiation in the Music Industry

Newly released music is never assessed in isolation by audiences, who tend to compare objective (sonic attributes) and subjective (genre affiliations) elements of its identity with the previous musical catalog of the corresponding artist, his or her musically proximal peers, and the most successful releases in the market. In this paper, we provide a general framework that disentangles the objective and subjective identities of new musical releases and evaluates how differentiation affects audience reactions. We posit that radio stations, who seek commercial success, have different preferences toward differentiation in comparison with critics, who grant cultural legitimacy in the industry. Combining play numbers, reviews, and music description data from different sources, we find that radio stations prefer consistency in the musical style of successive releases by the focal artist and especially favor albums that sound similar to chart toppers. Critics, however, exhibit opposite preferences, preferring novelty compared with past works of the artist, while remaining uninfluenced by contemporaneous music dynamics. Our moderation checks reveal that decisions around genre affiliations and scheduling of new releases aid music producers to effectively manage audience expectations. This suggests that data-driven decision support systems can help artists to strategically release new products that cater to heterogeneous tastes. This paper was accepted by David Simchi-Levi, entrepreneurship and innovation. Supplemental Material: The online appendix and data files are available at https://doi.org/10.1287/mnsc.2022.03319 .

The history of the formation and development of presentation for identification as a method of obtaining evidence in criminal proceedings

The scientific article examines the process of formation and transformation of the institution of presentation for identification in criminal proceedings, its place among investigative (detective) actions. It is emphasized that in the historical aspect, presentation for identification was inextricably linked with interrogation. In particular, it is noted that initially presentation for identification was considered as an integral part of interrogation, but over time it was realized that this procedure goes beyond its limits and should be considered as a separate investigative action. The basis for this was the establishment that the essence of identification consists in identifying a certain object by an imaginary image that was formed in a person’s memory in connection with the event under investigation. Thus, presentation for identification was formed as a type of investigative (search) action aimed at the direct use in proving images in the memory of a person who perceived the event of a criminal offense and individual objects. Its purpose is to establish that the object discovered during the investigation (person, thing, corpse) is the same one that was involved in the mechanism of committing a criminal offense, which is direct evidence in criminal proceedings. It is noted that the current CPC of Ukraine contains norms that regulate the presentation for identification of individual objects: a person (Article 228); things (Article 229); a corpse (Article 230). However, the norms of Articles 228-231 of the CPC of Ukraine contain mandatory requirements for preparation for conducting this investigative (search) action in different terms. In particular, they differ depending on the object of identification. In addition, one of the problematic issues that requires scientific research is the question of the correlation of the procedural procedure of presentation for identification with the tactical features of conducting this investigative (detective) action. Based on the historical analysis, a definition of the concept of presentation for identification is provided as an investigative (detective) action, which consists in identifying the presented object (person, thing, corpse) by an imaginary image preserved in the memory of a certain person.

Amyloid-β deposition in basal frontotemporal cortex is associated with selective disruption of temporal mnemonic discrimination.

Cerebral amyloid-beta (Aβ) accumulation, a hallmark pathology of Alzheimer's disease (AD), precedes clinical impairment by two to three decades. However, it is unclear whether Aβ contributes to subtle memory deficits observed during the preclinical stage. The heterogenous emergence of Aβ deposition may selectively impact certain memory domains, which rely on distinct underlying neural circuits. In this context, we tested whether specific domains of mnemonic discrimination, a neural computation essential for episodic memory, exhibit specific deficits related to early Aβ deposition. We tested 108 cognitively unimpaired human older adults (66% female) who underwent 18F-florbetapir positron emission tomography (Aβ-PET), and a control group of 35 young adults, on a suite of mnemonic discrimination tasks taxing object, spatial, and temporal domains. We hypothesized that Aβ pathology would be selectively associated with temporal discrimination performance due to Aβ's propensity to accumulate in the basal frontotemporal cortex, which supports temporal processing. Consistent with this hypothesis, we found a dissociation in which generalized age-related deficits were found for object and spatial mnemonic discrimination, while Aβ-PET levels were selectively associated with deficits in temporal mnemonic discrimination. Further, we found that higher Aβ-PET levels in medial orbitofrontal and inferior temporal cortex, regions supporting temporal processing, were associated with greater temporal mnemonic discrimination deficits, pointing to the selective vulnerability of circuits related to temporal processing early in AD progression. These results suggest that Aβ accumulation within basal frontotemporal regions may disrupt temporal mnemonic discrimination in preclinical AD, and future work is needed to determine whether assessing temporal mnemonic discrimination can aid in predicting emerging AD progression.Significance Statement Identifying subtle cognitive changes that reflect emerging Aβ accumulation could lead to the development of cognitive tasks to detect individuals at risk of AD and sensitively assess clinical intervention outcomes for Aβ-lowering therapeutics. Temporal mnemonic discrimination, the ability to distinguish between events occurring at different times, is supported by basal frontotemporal regions which are among the earliest impacted by Aβ. We demonstrate that Aβ-positivity, as well as increased Aβ deposition in basal frontotemporal regions, is selectively associated with deficits in temporal mnemonic discrimination, but not other domains of mnemonic discrimination such as object identity or spatial relationships. These results suggest that Aβ disrupts cortical circuits supporting temporal processing and highlights inclusion of temporal mnemonic discrimination tasks in future clinical assessments.

RODDMS-IoT: A Radar Based Real-Time Moving Object Detection and Distance Measurement System Using IoT for Enhanced Safety

The main users of RADAR (Radio Detection and Ranging) in its early development for security and monitoring purposes were the navy, the military, the aviation services, as well as the space organizations. Nowadays, the demand of radar is expanding. But it takes a while to detect, has a limited detection area, lacks target specificity due to its wide range, oversensitive, costly, etc. A cheaper, easy and effective alternate solution is to use ultrasonic sensor which use sound waves for detection and ranging. Therefore, this study offers a technique for using the Ultrasonic Sensor (HC-SR04) as RADAR. For the purpose of rotation and movement, the Servo Motor (SG90) is connected to the Ultrasonic Sensor (HCRS04). The SIM808 IoT integrated GPS-GSM module is also utilized to send out SMS notifications for object identification. These components are connected to an Arduino Uno and a Raspberry Pi3 to detect and notify the object. Additionally, the goal of this work is to capturing the distance, direction, and radar pulse duration and object shape simultaneously. An algorithm was used to determine the object's velocity, or its speed toward or away from the radar, in order to acquire the precision. By comparing the actual values with the data acquired by the radar, the accuracy of the measurement of both angle and distance is ensured. There are both metal and non-metal objects used in this study. This work's novelty lies in its ability to accurately detect real-time moving objects and measurement the distance of the objects. Besides proper distance and velocity, a large set of data was taken to find the accuracy of the radar for objects of different shapes. The results show the object detected with its distance and angle in a java based Graphical User Interface (GUI), different ranges of object in cm at which it is detected and the detection notification sent to the admin via SMS.

Benchmarking the robustness of the correct identification of flexible 3D objects using common machine learning models.

True three-dimensional (3D) data are prevalent in domains such as molecular science or computer vision. In these data, machine learning models are often asked to identify objects subject to intrinsic flexibility. Our study introduces two datasets from molecular science to assess the classification robustness of common model/feature combinations. Molecules are flexible, and shapes alone offer intra-class heterogeneities that yield a high risk for confusions. By blocking training and test sets to reduce overlap, we establish a baseline requiring the trained models to abstract from shape. As training data coverage grows, all tested architectures perform better on unseen data with reduced overfitting. Empirically, 2D embeddings of voxelized data produced the best-performing models. Evidently, both featurization and task-appropriate model design are of continued importance, the latter point reinforced by comparisons to recent, more specialized models. Finally, we show that the shape abstraction learned from database samples extends to samples that are evolving explicitly in time.

Remote Radio Frequency Sensing Based on 5G New Radio Positioning Reference Signals.

In this paper, the idea of a radar based on orthogonal frequency division multiplexing (OFDM) is applied to 5G NR Positioning Reference Signals (PRS). This study demonstrates how the estimation of the communication channel using the PRS can be applied for the identification of objects moving near the 5G NR receiver. In this context, this refers to a 5G NR base station capable of detecting a high-speed train (HST). The anatomy of a 5G NR frame as a sequence of OFDM symbols is presented, and different PRS configurations are described. It is shown that spectral analysis of time-varying channel impulse response weights, estimated with the help of PRS pilots, can be used for the detection of transmitted signal reflections from moving vehicles and the calculation of their time and frequency/Doppler shifts. Different PRS configurations with varying time and frequency reference signal densities are tested in simulations. The peak-to-noise-floor ratio (PNFR) of the calculated radar range-velocity maps (RVM) is used for quantitative comparison of PRS-based radar scenarios. Additionally, different echo signal strengths are simulated while also checking various observation window lengths (FFT lengths). This study proves the practicality of using PRS pilots in remote sensing; however, it shows that the most dense configurations do not provide notable improvements, while also demanding considerably more resources.

3D Terahertz Confocal Imaging with Chromatic Metasurface

AbstractTerahertz confocal imaging allows 3D see‐through of a non‐metallic object with high resolution. Conventional methods acquiring 3D images of thick objects suffer from limited depth‐of‐field, constrained depth resolution, and/or inconsistent spatial resolution at different depths. To address these limitations, the intrinsic chromatic aberration of a typical focusing metasurface is exploited to achieve frequency‐dependent focal lengths. An object located within this extended focal range can be readily 3D inspected by performing 2D raster scans. A rigorous analysis reveals that the focal spot maintains a constant waist diameter of 2.4 mm (equivalent to 2.2 at 275 GHz) and migrates 68.1 mm (equivalent to 62.4, or 16.4 times of Rayleigh length, or 1.4‐fold of the designed focal length at 275 GHz) from 175 to 525 GHz, and thus achieving a consistent spatial resolution and a large depth‐of‐field for 3D imaging. Importantly, this large depth‐of‐field is achieved with a relatively high numerical aperture of around 0.42. Measurements conducted between 220 and 330 GHz exhibit close agreement with the calculation. To demonstrate its imaging functionality, two stacked papers with different texts, a mobile phone, and earphones concealed in a charging case are imaged, where a short‐time Fourier transform is implemented in the time‐domain terahertz images to enhance image contrast. The presented metasurface is technologically significant for imaging systems to rapidly inspect objects in 3D with exceptional resolutions. Its potential applications include in‐situ defect detection and object identification in security screening.

Effect of roughness on object identification using OAM spectrum

Effect of roughness on object identification using OAM spectrum

CRFUSION: Fine-grained Object Identification Using RF-image Modality Fusion

CRFUSION: Fine-grained Object Identification Using RF-image Modality Fusion

From Fantômas to The Black Doll

American author and illustrator Edward Gorey (1925–2000) was open in interviews about the importance of silent cinema for his work, and in particular the films of the prolific French film director Louis Feuillade (1873–1925). Focusing on this underexplored aspect of Gorey’s work, the paper examines how Gorey employs visual, compositional, and narrative elements from Feuillade’s films, as well as their development of uncertainty within the areas of space, narrative, and character and object identity within the genre of nonsense. The influence of Feuillade is traced across such works as La Malle Saignante (The Bleeding Trunk, 1975), a bilingual homage to silent cinema, and Gorey’s silent screenplay The Black Doll (1973), while also identifying possible conventions of nonsense within Feuillade’s own silent films.

On the depth of identification of objects in the study of the influence of the density of the grid of wells on the degree of production of oil reserves

The article studies the effect of the well grid density on the final oil recovery coefficient under conditions of various groups of objects in carbonate reservoirs of the Volga-Ural oil and gas province. The issue being solved within the framework of this work is one of the most relevant and important when making various management decisions in conditions of uncertainty, significant variability and heterogeneity of geological and field data. More than 500 deposits of carbonate reservoirs were selected for the study, the development of which is carried out on the territory of the Ural-Volga region. To implement an integrated approach, the construction of models was carried out in several interrelated stages. Within the framework of basic modeling, field statistical methods were used to identify dependencies describing at a high quantitative and qualitative level the relationship between the oil recovery coefficient and the density of the well grid before flooding. To obtain the most representative results, a separate approach based on the preliminary division of objects into groups was also used. At the second stage, in addition, a deep differentiation of the initial objects was implemented to determine hidden patterns within already formed groups and a similar construction of models according to which interpretation was performed. When combining the results obtained, it was reliably established that models for assessing the degree of reserve production from the density of the well grid should be built not only in general by tectonic and stratigraphic elements, but also differentially by groups identified as a result of in-depth identification of development sites. It is proposed to use the obtained dependencies to solve the problems of optimizing the density of the well grid for both long-term oil deposits under development and newly commissioned and at the stage of drafting the first design documents. Keywords: well grid density; oil recovery coefficient; modeling; grouping of deposits of the Volga-Ural oil and gas province; deep identification of objects; carbonate reservoirs.

Modeling of the oil recovery process for managing the development of deposits using artificial intelligence

Based on a comprehensive study of the process of oil recovery from deposits in carbonate reservoirs of the Tournaisian tier of the Birskaya saddle, a forecast of the final coefficient of oil recovery for deposits that have been in development for a long time was carried out. The choice of the research object is due to the presence of a significant proportion of residual oil reserves, which can potentially be extracted as part of the implementation of an effective and sound asset management strategy. Previously, more than 60 objects were divided into four groups using one of the well-known methods of reducing the dimensionality of the problem ‒ the method of principal components, after which typical dependencies were established that make it possible to successfully group objects. Using statistical forecasting methods, models of the process of developing oil reserves were obtained differentially for various groups of selected objects, including a number of different parameters reflecting, among other things, geological heterogeneity. Based on the results of the interpretation of the obtained dependencies, the areas of their use in the plane of application of artificial intelligence systems in the development of deposits in the natural regime are established. Within the framework of determining the relationships between various parameters and evaluating the reliability of the obtained models, the need for object identification in solving a wide range of tasks to improve the efficiency of operation of both objects involved in the analysis and similar ones in terms of geological and commercial characteristics has been established. Keywords: oil recovery coefficient; oil field development; computer modeling; carbonate reservoirs of the Tournaisian tier of the Birskaya saddle; identification of objects.

Biomechanics based computer simulation of rural landscape design using remote sensing image technology

Introduction: The strategy and restoration of rural areas and landscape in biomechanics, with a particular emphasis on cellular and molecular biomechanics, is crucial for sustainable rural landscape design. At the cellular and molecular level, plants’ biomechanical properties, such as the rigidity and elasticity of cell walls, determine their growth patterns and responses to environmental factors. These properties are essential in understanding how plants can be effectively incorporated into the rural landscape to enhance its stability and functionality. Aim: The objective of this research is to develop a novel computer simulation model for rural landscape planning using remote sensing imaging technology. Research methodology: We introduce a novel Adaptive YOLOv7 method driven by Starling Murmuration search. UAVs are used to collect extensive visual data for training the model. By considering cellular and molecular biomechanics, we can analyze how the mechanical forces within plants affect their ability to resist wind, retain water, and interact with the surrounding soil and other organisms. This knowledge can be integrated into the model to better predict the long-term viability and adaptability of different plant species in the rural landscape. The combination of the 3D GIS virtual image strategy model and our proposed model, along with SM optimization, not only improves object identification but also takes into account the biomechanical aspects for more accurate simulations. Crowdsourcing helps in precisely mapping rural landscapes and structures, while the incorporation of biomechanical principles ensures better adaptability to changing environmental and ecological conditions. Findings and Conclusion: Implemented in Python software, our SM-AYOLOv7 model shows excellent performance, with metrics like f1 score (93.64%), recall (92.34%), accuracy (91.72%), and IoU (90.23%). Our method outperforms conventional ones, demonstrating enhanced accuracy and flexibility, especially in handling changing configurations, due to the integration of cellular and molecular biomechanical insights.

Semantic segmentation of urban airborne LiDAR data of varying landcover diversity using XGBoost

AbstractSemantic segmentation of aerial LiDAR dataset is a crucial step for accurate identification of urban objects for various applications pertaining to sustainable urban development. However, this task becomes more complex in urban areas characterised by the coexistence of modern developments and natural vegetation. The unstructured nature of point cloud data, along with data sparsity, irregular point distribution, and varying sizes of urban objects, presents challenges in point cloud classification. To address these challenges, development of robust algorithmic approach encompassing efficient feature sets and classification model are essential. This study incorporates point‐wise features to capture the local spatial context of points in datasets. Furthermore, an ensemble machine learning model based on extreme boosting is utilised, which integrates sequential training for weak learners, to enhance the model’s resilience. To thoroughly investigate the efficacy of the proposed approach, this study utilises three distinct datasets from diverse geographical locations, each presenting unique challenges related to class distribution, 3D terrain intricacies, and geographical variations. The Land‐cover Diversity Index is introduced to quantify the complexity of landcover in 3D by measuring the degree of class heterogeneity and the frequency of class variation in the dataset. The proposed approach achieved an accuracy of 90% on the regionally complex, higher landcover diversity dataset, Trivandrum Aerial LiDAR Dataset. Furthermore, the results of the study demonstrate improved overall predictive accuracy of 91% and 87% on data segments from two benchmark datasets, DALES and Vaihingen 3D.