Motion Sequences Research Articles

AMD: Anatomical Motion Diffusion with Interpretable Motion Decomposition and Fusion

Generating realistic human motion sequences from text descriptions is a challenging task that requires capturing the rich expressiveness of both natural language and human motion. Recent advances in diffusion models have enabled significant progress in human motion synthesis. However, existing methods struggle to handle text inputs that describe complex or long motions. In this paper, we propose the Adaptable Motion Diffusion (AMD) model, which leverages a Large Language Model (LLM) to parse the input text into a sequence of concise and interpretable anatomical scripts that correspond to the target motion. This process exploits the LLM’s ability to provide anatomical guidance for complex motion synthesis. We then devise a two-branch fusion scheme that balances the influence of the input text and the anatomical scripts on the inverse diffusion process, which adaptively ensures the semantic fidelity and diversity of the synthesized motion. Our method can effectively handle texts with complex or long motion descriptions, where existing methods often fail. Experiments on datasets with relatively more complex motions, such as CLCD1 and CLCD2, demonstrate that our AMD significantly outperforms existing state-of-the-art models.

Existence Is Chaos: Enhancing 3D Human Motion Prediction with Uncertainty Consideration

Human motion prediction is consisting in forecasting future body poses from historically observed sequences. It is a longstanding challenge due to motion's complex dynamics and uncertainty. Existing methods focus on building up complicated neural networks to model the motion dynamics. The predicted results are required to be strictly similar to the training samples with L2 loss in current training pipeline. However, little attention has been paid to the uncertainty property which is crucial to the prediction task. We argue that the recorded motion in training data could be an observation of possible future, rather than a predetermined result. In addition, existing works calculate the predicted error on each future frame equally during training, while recent work indicated that different frames could play different roles. In this work, a novel computationally efficient encoder-decoder model with uncertainty consideration is proposed, which could learn proper characteristics for future frames by a dynamic function. Experimental results on benchmark datasets demonstrate that our uncertainty consideration approach has obvious advantages both in quantity and quality. Moreover, the proposed method could produce motion sequences with much better quality that avoids the intractable shaking artefacts. We believe our work could provide a novel perspective to consider the uncertainty quality for the general motion prediction task and encourage the studies in this field. The code will be available in https://github.com/Motionpre/Adaptive-Salient-Loss-SAGGB.

MotionMix: Weakly-Supervised Diffusion for Controllable Motion Generation

Controllable generation of 3D human motions becomes an important topic as the world embraces digital transformation. Existing works, though making promising progress with the advent of diffusion models, heavily rely on meticulously captured and annotated (e.g., text) high-quality motion corpus, a resource-intensive endeavor in the real world. This motivates our proposed MotionMix, a simple yet effective weakly-supervised diffusion model that leverages both noisy and unannotated motion sequences. Specifically, we separate the denoising objectives of a diffusion model into two stages: obtaining conditional rough motion approximations in the initial T-T* steps by learning the noisy annotated motions, followed by the unconditional refinement of these preliminary motions during the last T* steps using unannotated motions. Notably, though learning from two sources of imperfect data, our model does not compromise motion generation quality compared to fully supervised approaches that access gold data. Extensive experiments on several benchmarks demonstrate that our MotionMix, as a versatile framework, consistently achieves state-of-the-art performances on text-to-motion, action-to-motion, and music-to-dance tasks.

AMD: Autoregressive Motion Diffusion

Human motion generation aims to produce plausible human motion sequences according to various conditional inputs, such as text or audio. Despite the feasibility of existing methods in generating motion based on short prompts and simple motion patterns, they encounter difficulties when dealing with long prompts or complex motions. The challenges are two-fold: 1) the scarcity of human motion-captured data for long prompts and complex motions. 2) the high diversity of human motions in the temporal domain and the substantial divergence of distributions from conditional modalities, leading to a many-to-many mapping problem when generating motion with complex and long texts. In this work, we address these gaps by 1) elaborating the first dataset pairing long textual descriptions and 3D complex motions (HumanLong3D), and 2) proposing an autoregressive motion diffusion model (AMD). Specifically, AMD integrates the text prompt at the current timestep with the text prompt and action sequences at the previous timestep as conditional information to predict the current action sequences in an iterative manner. Furthermore, we present its generalization for X-to-Motion with “No Modality Left Behind”, enabling for the first time the generation of high-definition and high-fidelity human motions based on user-defined modality input.

Explore 3D Dance Generation via Reward Model from Automatically-Ranked Demonstrations

This paper presents an Exploratory 3D Dance generation framework, E3D2, designed to address the exploration capability deficiency in existing music-conditioned 3D dance generation models. Current models often generate monotonous and simplistic dance sequences that misalign with human preferences because they lack exploration capabilities.The E3D2 framework involves a reward model trained from automatically-ranked dance demonstrations, which then guides the reinforcement learning process. This approach encourages the agent to explore and generate high quality and diverse dance movement sequences. The soundness of the reward model is both theoretically and experimentally validated. Empirical experiments demonstrate the effectiveness of E3D2 on the AIST++ dataset.

The Algorithms for Managing a Matrix-Based Warehouse Utilizing Standardized Transport and Storage Cells

The article explores the principles of organizing warehouse systems based on standardized transport and storage cells. It discusses the general hardware features of transport and storage cells that enable the movement of stored goods between them. The article presents a method for constructing a warehouse graph that takes into account possible directions for transferring containers with cargo between the transport and storage cells integrated into the warehouse structure, along with their common hardware characteristics. Key criteria used in determining the edge weights of the graph are described, including the basic cost of movement for each axis, cell wear and tear, cargo weight, fragility of the cargo, distance to the nearest available cells, and repairability. Algorithms are presented, the primary task of which is to determine the sequence of container movements between warehouse cells to facilitate loading and unloading operations. Simulation modeling of a warehouse with dimensions of 5×5×5 was conducted using the proposed algorithms, both with and without considering cell wear and tear parameters. The results of the simulation highlighted the significance of this criterion, allowing for extended warehouse servicing intervals and maximizing the time until the first failure.The study also investigates the possibility of optimizing the structure of such warehouse systems to meet various requirements. As part of this investigation, the structure was optimized for a warehouse with dimensions of 4×3×3.

Disc Degeneration and Cervical Spine Intervertebral Motion: A Cross-Sectional Study in Patients with Neck Pain and Matched Healthy Controls.

While neck pain can be defined in clinical terms, in most cases the underlying pathophysiology is largely unknown. Regional cervical spine range of motion is often found to be reduced in patients with neck pain compared to persons without pain although it is not clear if the decreased range is cause or effect. Less is known about the role of intervertebral kinematics and how that might be related to the presence of disc degeneration. In this study, the prevalence of intervertebral disc degeneration and continuous cervical intervertebral motion were both measured utilizing quantitative fluoroscopy (QF) in patients with subacute or chronic neck pain (n = 29) and gender-matched healthy controls (n = 30). A composite disc degeneration (CDD) score was calculated for each participant from the first, neutral, lateral fluoroscopic image. Intervertebral motion sharing parameters of motion-sharing inequality (MSI) and motion-sharing variability (MSV) were derived from the active cervical motion sequences obtained while patients were seated. The objective was to determine if average age, CDD, MSI, and MSV values were correlated and if there were differences in these variables between the neck pain group and the healthy control group. Correlation analysis was conducted for age, CDD, MSI, and MSV in each group. Age was moderately correlated with MSV in cervical spine extension in patients only (r = 0.63, p < 0.001). There were no significant differences in the prevalence of disc degeneration (CDD) between patients, who had on average mild pain and related disability, and healthy controls (median CDD 2 both groups, p = 0.94). There were also no significant differences in either flexion or extension intervertebral motion-sharing inequality or variability (MSI or MSV) between groups as measured during active cervical motion.

Prediction of cognitive conflict during unexpected robot behavior under different mental workload conditions in a physical human–robot collaboration

Objective. Brain–computer interface (BCI) technology is poised to play a prominent role in modern work environments, especially a collaborative environment where humans and machines work in close proximity, often with physical contact. In a physical human robot collaboration (pHRC), the robot performs complex motion sequences. Any unexpected robot behavior or faulty interaction might raise safety concerns. Error-related potentials, naturally generated by the brain when a human partner perceives an error, have been extensively employed in BCI as implicit human feedback to adapt robot behavior to facilitate a safe and intuitive interaction. However, the integration of BCI technology with error-related potential for robot control demands failure-free integration of highly uncertain electroencephalography (EEG) signals, particularly influenced by the physical and cognitive state of the user. As a higher workload on the user compromises their access to cognitive resources needed for error awareness, it is crucial to study how mental workload variations impact the error awareness as it might raise safety concerns in pHRC. In this study, we aim to study how cognitive workload affects the error awareness of a human user engaged in a pHRC. Approach. We designed a blasting task with an abrasive industrial robot and manipulated the mental workload with a secondary arithmetic task of varying difficulty. EEG data, perceived workload, task and physical performance were recorded from 24 participants moving the robot arm. The error condition was achieved by the unexpected stopping of the robot in 33% of trials. Main results. We observed a diminished amplitude for the prediction error negativity (PEN) and error positivity (Pe), indicating reduced error awareness with increasing mental workload. We further observed an increased frontal theta power and increasing trend in the central alpha and central beta power after the unexpected robot stopping compared to when the robot stopped correctly at the target. We also demonstrate that a popular convolution neural network model, EEGNet, could predict the amplitudes of PEN and Pe from the EEG data prior to the error. Significance. This prediction model could be instrumental in developing an online prediction model that could forewarn the system and operators of the diminished error awareness of the user, alluding to a potential safety breach in error-related potential-based BCI system for pHRC. Therefore, our work paves the way for embracing BCI technology in pHRC to optimally adapt the robot behavior for personalized user experience using real-time brain activity, enriching the quality of the interaction.

Age-related changes in motor planning for prior intentions: a mouse tracking reach-to-click task.

When we complete sequential movements with different intentions, we plan our movements and adjust ahead. Such a phenomenon is called anticipatory planning for prior intentions and is known to decline with age. In daily life activities, we often need to consider and plan for multiple demands in one movement sequence. However, previous studies only considered one dimension of prior intentions, either different types of onward actions or different precisions of fit or placement. Therefore, in this study, we investigated anticipatory planning for both extrinsic (movement direction) and intrinsic (fit precision) target-related properties in a computer-based movement task and analyzed the computer cursor movement kinematics of both young and older adults. We found that older people consider and adjust for different properties step-by-step, with movement direction being considered as a prior intention during reach movement and fit precision as a motor constraint during drop movement. The age-related changes in the completion of onward actions are constrained by one's general cognitive ability, sensorimotor performance and effective motor planning for prior intentions. Age-related decline in motor planning can manifest as counterproductive movement profiles, resulting in suboptimal performance of intended actions.

Dual-Task and Single-Task Practice Does Not Influence the Attentional Demands of Movement Sequence Representations

This study examined the attentional demands of movement sequence representations at different temporal points after single- or dual-task practice. The visual-spatial representation encodes the movement based on visual-spatial coordinates such as the target locations. The motor representation encodes the movement in motor coordinates including joint angles and muscle activation patterns. Participants were randomly assigned to a single-task or dual-task practice group. Following acquisition, participants performed two retention tests and inter-manual transfer tests, both under dual-task and single-task. The transfer tests consisted of a mirror and non-mirror test and examined motor and visual-spatial representation development. The main finding is that attentional demands of the sequence representations were not affected by the practice condition. However, movement initiation requires more attention than the end of the movement in both representations.

Wiring connector-terminated cables based on manipulation planning with collision-free EMD net

In this paper, we propose a manipulation planning method for cable wiring in the assembly of electric appliances etc. We address a scenario where a robot grasps a connector attached to the end of a cable and has to bring the connector to a socket. To accomplish this automatically, we propose a novel manipulation planning method. The method extends the Encode-Manipulate-Decode network (EMD net), which can predict shape changes of deformable objects and generate robot motion sequences for producing desired shape transitions. This enables us to find connector trajectories that avoid collision between the cable and the surrounding environment. We conducted experiments with several different cable lengths. We also introduce some functions required for real-world wiring, such as online cable shape modification. Experimental results show that the proposed method can achieve stable manipulation of real cables.

Детализация образа как показатель образно-пластического творчества дошкольников

&lt;p&gt;Based on a brief description of creativity and imagination, their indicators and age characteristics, the insufficiency of research on the creativity of preschoolers is shown. A diagnostic study (1993&amp;ndash;2000) is described, aimed at identifying the age characteristics of children 3&amp;ndash;7 years old in detailing the motor-plastic image. 192 children participated in the diagnosis according to the author's method "Hare and Wolf"; they were asked to embody the plot of the "fairy tale" by means of the movement language. It was revealed: in the younger group, schematic movements predominate, signifying the image; in the middle group, the proportion of "complex" figurative movements is growing; in the senior and preparatory groups, "complexes" (partial and integral) prevail and the number of sequences of figurative movements that convey the phases of the character's actions is growing. Statistical analysis (according to the criteria H Kruskal-Wallis and U Mann-Whitney) revealed significant differences in image detail between different age samples and the absence of such differences between boys and girls (in different years).&lt;/p&gt;

GAZE DIRECTION MONITORING MODEL IN COMPUTER SYSTEM FOR ACADEMIC PERFORMANCE ASSESSMENT

This paper focuses on the research and application of eye movement and gaze direction analysis in online testing systems. The practical novelty of the proposed model of monitoring the direction of gaze in a computer-based knowledge control system lies in the possibility of automated remote control over a large audience of students. The practical significance lies in creating the same conditions for computer testing for all students and increasing the correspondence of knowledge level to the received test results. The implemented and tested system is relevant and necessary in higher educational institutions, particularly in Ukraine, where remote education has emerged as the safest means of acquiring knowledge. This is especially true for fields of study where practical tasks and laboratory work do not necessitate a student's physical presence at the institution. That is why the application of the latest information technologies is extremely important. The dynamic authentication based on the sequence of eye movements proposed in the model allows error-free user's eye area detection for further analysis of the test subject's behavior. The proposed authentication method excludes the need to enter passwords or CAPTCHAs, ensures the speed of determining the user's presence. Further analysis of the user's direction of vision includes responding to the information received, such as skipping a question or the need for re-authorization. Question skipping occurs when the system decides that the user has not been looking at the screen for an extended period (looking sideways, down, up for more than 30 seconds). Re-authentication becomes necessary if the user exits the test or if there is a user replacement. Real-time gaze control capability is implemented by using massive parallel processing system (NVIDIA GeForce GTX 1650 graphics card) for calculations. The analysis of the results obtained shows that the proposed approach allows gaze detection at different illumination ranges (from 0.3 lux to 10,000 lux), as well as to detect states of the user's eyes that violating test rules for further system response (skipping a question or re-authentication request).

Open Dataset for Testing of Visual SLAM Algorithms under Different Weather Conditions

Existing datasets for testing SLAM algorithms in outdoor environments are not suitable for assessing the influence of weather conditions on localization accuracy. Obtaining a suitable dataset from the real world is difficult due to the long data collection period and the inability to exclude dynamic environmental factors. Artificially generated datasets make it possible to bypass the described limitations, but up to date, researchers have not identified testing SLAM algorithms under different weather conditions as a stand-alone task, despite the fact that it is one of the main aspects of the difference between outdoor and indoor environments. This work presents a new open dataset that consists of 36 sequences of robot movement in an urban environment or rough terrain, in the form of images from a stereo camera and the ground truth position of the robot, collected at a frequency of 30 Hz. Movement within one area occurs along a fixed route; the sequences are distinguished only by whether conditions, which can make it possible to correctly assess the influence of weather phenomena on the accuracy of localization.

Sensor based Dance Coherent Action Generation Model using Deep Learning Framework

Dance Coherent Action Generation is a popular research task in recent years to generate movements and actions for computer-generated characters in a simulated environment. It is sometimes referred to as "Motion Synthesis". Motion synthesis algorithms are used to generate physically believable, visually compelling, and contextually appropriate movement using motion sensors. The Dance Coherent Action Generation Model (DCAM) is a generative framework for producing aesthetically pleasing movements using deep learning from small amounts of data. By learning an internal representation of motion dynamics, DCAM can synthesize long sequences of movements in which coherent patterns can be created through latent space interpolation. This framework provides a mechanism for varying the amplitude of the generated motion, allowing further realistic thinking and expression. The proposed model obtained 93.79% accuracy, 93.79% precision, 97.75% recall and 92.92% F1 score. DCAM exploits the balance between imitation and creativity by enabling the production of novel outputs from limited input data and can be trained in an unsupervised manner or fine-tuned with sparse supervision. Furthermore, the framework is easily extended to handle multiple layers of abstraction and can be further personalized to a particular type of movement, enabling the generation of highly individualized outputs.

Dopamine neuron activity encodes the length of upcoming contralateral movement sequences

Dopaminergic neurons (DANs) in the substantia nigra pars compacta (SNc) have been related to movement speed, and loss of these neurons leads to bradykinesia in Parkinson's disease (PD). However, other aspects of movement vigor are also affected in PD; for example, movement sequences are typically shorter. However, the relationship between the activity of DANs and the length of movement sequences is unknown. We imaged activity of SNc DANs in mice trained in a freely moving operant task, which relies on individual forelimb sequences. We uncovered a similar proportion of SNc DANs increasing their activity before either ipsilateral or contralateral sequences. However, the magnitude of this activity was higher for contralateral actions and was related to contralateral but not ipsilateral sequence length. In contrast, the activity of reward-modulated DANs, largely distinct from those modulated by movement, was not lateralized. Finally, unilateral dopamine depletion impaired contralateral, but not ipsilateral, sequence length. These results indicate that movement-initiation DANs encode more than a general motivation signal and invigorate aspects of contralateral movements.

Dancing without seeing: The influence of vision on movement sequence learning and motor imagery in dance

ABSTRACT Dance is commonly taught through model observation, aided by verbal and gestural cues. Trained dancers are also capable of learning dance movements from verbal instructions in the absence of a visual model, using elaborate multimodal movement representations in long-term memory acquired through training. The role of vision in this situation remains unclear. To investigate the influence of vision on dance sequence learning from verbal instruction, we conducted an online study with 76 professional-level dancers. Participants filmed themselves while learning and subsequently recalling and reproducing a dance phrase from spoken verbal instructions presented as pre-recorded audio. They were randomly assigned to one of two conditions: closing their eyes and wearing a blindfold or keeping their eyes open during learning. As a measure of learning success, recall performance was evaluated with quantitative movement criteria. Motor imagery during learning and recall was assessed via self-report questionnaires. Results showed no differences between the two conditions in recall performance or motor imagery; however, external and kinaesthetic motor imagery quality and quantity showed large to medium positive correlations with recall performance. The findings suggest that vision may not be crucial for successful learning of movement sequences from verbal instructions alone. Motor imagery, however, could be a powerful tool in dance learning as it may enhance memory for movement sequences. Further research should focus on the use of imagery (especially dynamic imagery) in dance during different phases of movement learning and recall under varying sensory and task-based (artistic) conditions.

Further experiments and remarks regarding the possible formation of blood stains on the Turin Shroud: stains attributed to the nailing of the hands

The formation of red discolorations (‘blood stains’) on the Turin Shroud (TS), a Christian relic believed to be the burial cloth of Jesus of Nazareth, is controversially discussed. We performed experiments to identify possible explanations for the formation of the stains on the hands and forearms of the Turin Shroud Man (TSM). In preliminary non-standardised experiments, after applying blood to the dorsal and palmar side of the probands’ wrists, they moved their arms around at their own discretion to provoke blood flows as similar as possible to those on the TS. A blood stain pattern similar to that on the left wrist could be provoked by slowly turning the wrist to the ulnar side. In contrast, a branched pattern of multiple streaks, as depicted on the forearms, was difficult to reproduce. In a standardised test setup, the probands moved their dry, dirtied, or oiled arms jerkily in a predetermined sequence of movements. More body hair only slightly facilitated the formation of a branched pattern. On oiled skin, however, the formation of branches was significantly facilitated. This may support the hypothesis that the blood stains on the forearms were formed by moving the body between the unnailing and the burial. The formation of a branched pattern seems feasible if the arms were moved jerkily and were possibly exposed to water and oils postmortem (e.g. transporting the washed and oiled body). Nevertheless, the well-defined blood stains with multiple branchings are difficult to explain. Additionally, the blood stains on the forearms may have originated from deep scourging wounds, where dried blood was again mobilised by water (and oil). We are aware that no reliable conclusions about the formation of the ‘blood stains’ on the TS can be drawn from our findings. However, they may contribute to the discussion on this topic.

Velocity-aware spatial-temporal attention LSTM model for inverse dynamic model learning of manipulators.

An accurate inverse dynamics model of manipulators can be effectively learned using neural networks. However, further research is required to investigate the impact of spatiotemporal variations in manipulator motion sequences on network learning. In this work, the Velocity Aware Spatial-Temporal Attention Residual LSTM neural network (VA-STA-ResLSTM) is proposed to learn a more accurate inverse dynamics model, which uses a velocity-aware spatial-temporal attention mechanism to extract dynamic spatiotemporal features selectively from the motion sequence of the serial manipulator. The multi-layer perception (MLP) attention mechanism is adopted to capture the correlation between joint position and velocity in the motion sequence, and the state correlation between hidden units in the LSTM network to reduce the weight of invalid features. A velocity-aware state fusion approach of LSTM network hidden units' states is proposed, which utilizes variation in joint velocity to adapt to the temporal characteristics of the manipulator dynamic motion, improving the generalization and accuracy of the neural network. Comparative experiments have been conducted on two open datasets and a self-built dataset. Specifically, the proposed method achieved an average accuracy improvement of 61.88% and 43.93% on the two different open datasets and 71.13% on the self-built dataset compared to the LSTM network. These results demonstrate a significant advancement in accuracy for the proposed method. Compared with the state-of-the-art inverse dynamics model learning methods of manipulators, the modeling accuracy of the proposed method in this paper is higher by an average of 10%. Finally, by visualizing attention weights to explain the training procedure, it was found that dynamic modeling only relies on partial features, which is meaningful for future optimization of inverse dynamic model learning methods.

Automated Discovery of Successful Strategies in Association Football

Using automated data analysis to understand what makes a play successful in football can enable teams to make data-driven decisions that may enhance their performance throughout the season. Analyzing different types of plays (e.g., corner, penalty, free kicks) requires different considerations. This work focuses on the analysis of corner kick plays. However, the central ideas apply to analyzing all types of plays. While prior analyses (univariate, bivariate, multivariate) have explored the link between contextual factors (e.g., match period, type of defensive marking) and the level of success of a corner kick (e.g., shot, shot on goal, goal), there has been no attempt to combine spatiotemporal event data (sequences of ball movements through the field) and contextual information to determine when and how (strategy) a particular type of corner kick play (tactic) is more likely to succeed or not. To address this gap, we propose an approach that (1) transforms spatiotemporal data into an alternative representation suitable for mining sequential patterns, (2) identifies and characterizes the sequential patterns used by offensive teams to move the ball toward the scoring zone (tactics), and (3) extracts contrast patterns to identify under what conditions different tactics result in increased chances of success or failure; we call these conditions strategies. Our results suggest that favorable and unfavorable conditions for tactic application are not the same across different tactics, supporting the argument that there is a benefit in performing an analysis that treats different tactics separately, where spatiotemporal information plays a crucial role. Unlike prior works on the corner kick, our approach can capture how the interaction between multiple contextual factors impacts the outcome of a corner kick. At the same time, the results can be explained to others in natural languages.