Final Action Research Articles

Determination of Lactose in Lactose-Free and Low-Lactose Milk, Milk Products, and Products Containing Dairy Ingredients by the LactoSens®R Amperometry Method: Final Action 2020.01.

The LactoSens®R method was previously shown to have acceptable accuracy and repeatability precision as required by AOAC Standard Method Performance Requirements (SMPR®) 2018.009 for determination of lactose in low-lactose or lactose-free milk, milk products, and products containing dairy ingredients and was awarded Official Method of AnalysisSM (OMA) First Action status in 2020. The method was subjected to a multilaboratory validation (MLV) study to evaluate the reproducibility precision of the method. Fourteen validation materials were provided to 15 laboratories in seven countries as blind duplicates. The materials ranged from 0 to 173 mg/100 g lactose. Each laboratory analyzed the blind duplicates according to OMA 2020.01. The data were analyzed for repeatability and reproducibility precision. RSDr values varied from 2.81 to 8.76%, and RSDR values varied from 4.25 to 12.5%. When sorted by category and concentration range, these results met the repeatability and reproducibility criteria required by SMPR 2018.009. The data generated in the MLV support the adoption of OMA 2020.01 as Final Action status. The LactoSensR method, as described by OMA 2020.01, provides an accurate and precise determination of lactose in a variety of low-lactose and lactose-free milk, milk products, and products containing dairy ingredients in minutes.

The characteristics of counterattacks in the Spanish league (La Liga) in 2021-2022.

Current studies on counterattacks are limited despite the fact that it is one of the most effective ways to score in soccer. The aim of the study was to investigate the characteristics of counterattacks in the Spanish league (La Liga) for the year 2021-2022. The sample used in the study were 2631 counterattacks collected from a total of 380 games. The variables used were: 1) the space which the counterattack had started; 2) the number of players who participated; 3) the number of defending players; 4) the duration of the counterattack; 5) the touches to complete the counterattack; 6) the formation of both teams; 7) the way that counterattack was completed (final action); 8) the minute that the counterattack took place; and 9) which team had a home advantage. The results were analyzed using the Chi-square test. The level of significance was defined at P<0.05. Firstly, it appeared that most counterattacks began from the axis of the defensive zone lasting 6-15 seconds. There were 3-4 attackers participating against 4-6 defenders, using 5-9 touches. In addition, most of them took place in the 46-60 minute of the game and were completed in a shot. The study can help coaches and analysts to understand the characteristics of the counterattacks and adapt their training in a way to make their teams more effective, both in the attacking and defending phase.

Emotional Contagion-Aware Deep Reinforcement Learning for Antagonistic Crowd Simulation

The antagonistic behavior in the crowd usually exacerbates the seriousness of the situation in sudden riots, where the antagonistic emotional contagion and behavioral decision making play very important roles. However, the complex mechanism of antagonistic emotion influencing decision making, especially in the environment of sudden confrontation, has not yet been explored very clearly. In this paper, we propose an Emotional contagion-aware Deep reinforcement learning model for Antagonistic Crowd Simulation (ACSED). Firstly, we build a group emotional contagion module based on the improved Susceptible Infected Susceptible (SIS) infection disease model, and estimate the emotional state of the group at each time step during the simulation. Then, the tendency of crowd antagonistic action is estimated based on Deep Q Network (DQN), where the agent learns the action autonomously, and leverages the mean field theory to quickly calculate the influence of other surrounding individuals on the central one. Finally, the rationality of the predicted actions by DQN is further analyzed in combination with group emotion, and the final action of the agent is determined. The proposed method in this paper is verified through several experiments with different settings. We can conclude antagonistic emotions play a critical role in the decision making of the crowd through influencing the individual behavior in the riot scenario, where individual behaviors are primarily driven by emotions and goals, rather than common rules. The experiment results also prove that the antagonistic emotion has a vital impact on the group combat, and positive emotional states are more conducive to combat. Moreover, by comparing the simulation results with real scenes, the feasibility of our method is further confirmed, which can provide good reference to formulate battle plans and improve the win rate of righteous groups in a variety of situations.

Study of the Effect of Some Muzzle Device Types on the Firing Force and Firing Impulse

In this article, a mathematical model is used to determine the force and impulse of the weapon’s firing. The force and impulse of the weapon’s firing are determined during the whole firing process, which consists of two stages: the stage projectile moving inside the barrel and the stage at the final action stage of combustion gas. The input parameters of the UK-59 machine gun in the Czech Republic were selected to verify the mathematical model. The results of the article are the law of combustion gas pressure in the barrel, the law of combustion gas pressure in the gas chamber, and the law of force and impulse of the shot corresponding to different types of muzzle devices. This result is consistent with the experimental results presented in the document, and the error is less than 3%. Besides, the article also analyzes the influence of some muzzle device types on the firing force and firing impulse. The method of numerical integration has been applied to solve the problem using Matlab software. The research results of the article are a reliable theoretical basis for determining the firing forces, serving the problem of designing the mount as well as the firing stability of the weapon when firing. This is the basis for the optimal selection and design of the muzzle device used in the weapon.

A general skeleton-based action and gesture recognition framework for human–robot collaboration

Recognizing human actions is crucial for an effective and safe collaboration between humans and robots. For example, in a collaborative assembly task, human workers can use gestures to communicate with the robot, and the robot can use the recognized actions to anticipate the next steps in the assembly process, leading to improved safety and productivity. In this work, we propose a general framework for human action recognition based on 3D pose estimation and ensemble techniques, which allows to recognize both body actions and hand gestures. The framework relies on OpenPose and 2D to 3D lifting methods to estimate 3D joints for the human body and the hands, feeding then these joints into a set of graph convolutional networks based on the Shift-GCN architecture. The output scores of all networks are combined using an ensemble approach to predict the final human action. The proposed framework was evaluated on a custom dataset designed for human–robot collaboration tasks, named IAS-Lab Collaborative HAR dataset. The results showed that using an ensemble of action recognition models improves the accuracy and robustness of the overall system; moreover, the proposed framework can be easily specialized on different scenarios and achieve state-of-the-art results on the HRI30 dataset when coupled with an object detector or classifier.

Wearable rehabilitation wristband for distal radius fractures.

Distal radius fractures are a common type of fracture. For patients treated with closed reduction with splinting, a period of rehabilitation is still required after the removal of the splint. However, there is a general lack of attention and low compliance to rehabilitation training during this period, so it is necessary to build a rehabilitation training monitoring system to improve the efficiency of patients' rehabilitation. A wearable rehabilitation training wristband was proposed, which could be used in the patient's daily rehabilitation training scenario and could recognize four common wrist rehabilitation actions in real-time by using three thin film pressure sensors to detect the pressure change curve at three points on the wrist. An algorithmic framework for classifying rehabilitation training actions was proposed. In our framework, an action pre-detection strategy was designed to exclude false detections caused by switching initial gestures during rehabilitation training and wait for the arrival of the complete signal. To classify the action signals into four categories, firstly an autoencoder was used to downscale the original signal. Six SVMs were then used for evaluation and voting, and the final action with the highest number of votes would be used as the prediction result. Experimental results showed that the proposed algorithmic framework achieved an average recognition accuracy of 89.62%, an average recognition recall of 88.93%, and an f1 score of 89.27% on the four rehabilitation training actions. The developed device has the advantages of being small size and easy to wear, which can quickly and accurately identify and classify four common rehabilitation training actions. It can easily be combined with peripheral devices and technologies (e.g., cell phones, computers, Internet) to build different rehabilitation training scenarios, making it worthwhile to use and promote in clinical settings.

Solving Rearrangement Puzzles Using Path Defragmentation in Factored State Spaces

Rearrangement puzzles are variations of rearrangement problems in which the elements of a problem are potentially logically linked together. To efficiently solve such puzzles, we develop a motion planning approach based on a new state space that is logically <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">factored</i> , integrating the capabilities of the robot through factors of simultaneously manipulatable joints of an object. Based on this factored state space, we propose less-actions RRT (LA-RRT), a planner which optimizes for a low number of actions to solve a puzzle. At the core of our approach lies a new path defragmentation method, which rearranges and optimizes consecutive edges to minimize action cost. We solve six rearrangement scenarios with a Fetch robot, involving planar table puzzles and an escape room scenario. LA-RRT significantly outperforms the next best asymptotically-optimal planner by 4.01 to 6.58 times improvement in final action cost.

РАЗВИТИЕ ВНЕШНЕЭКОНОМИЧЕСКОЙ ДЕЯТЕЛЬНОСТИ ОТРАСЛЕВОГО СЕКТОРА РЕГИОНА

One of the key issues for the effective development of the agricultural sector of the economy is foreign economic activity. The study was conducted in order to study the level of development of the sectoral market of certain types of agricultural products in an agro-oriented cross-border region in the context of reaching the international level and the availability of logistics infrastructure. The indicators of the promising leading sectors of the agricultural sector of the economy (economic, innovative, infrastructural) considered for evaluation make it possible to objectively assess the level of sectoral development and give appropriate recommendations, taking into account the logistical capabilities of the region and its cross-border location, the effective functioning of crop and livestock industries in the context of the development of international activity. Effective sectoral development involves the interaction of all constituent elements, while a special role is assigned to the processing link, which has the final action in the technological production chain in the form of the presentation of finished products on the consumer market, as well as the logistics infrastructure, which allows optimizing transportation costs. The proposed segmentation of local production areas on the example of the dairy industry based on the potential in the form of production of raw milk and its processing suggests the presence of a cheese-making zone and a zone for the production of whole-milk products, oriented to the domestic and international markets. The developed transport infrastructure of the region and the cross-border position allows integrating into international transport corridors, which will allow more efficient development of sectoral markets for agricultural products. The results of the study can be the basis for making a forecast for the sustainable development of the agricultural sectors of the region, the processing industry in the context of active foreign economic activity.

Correlations of passes and playing formations with technical-tactical elements during the 2022 FIFA World Cup.

This study on the 2022 Qatar FIFA World Cup examined: a) the existence of correlations between passes with other technical-tactical elements, b) the differences in technical-tactical elements in relation to team formation, and c) the characteristics of the key passes created during games. Sixty-four games were analyzed. Match-analysis data were collected using InStat database. Spearman correlation revealed that the number of passes and key passes correlated with number of goals, chances and shots (P<0.05). Team ball possession was correlated with number of passes, key passes, chances, shots and the organized attacks (P<0.001). Kruskal-Wallis H non-parametric test revealed that the teams with the higher percentages of team ball possession, and the higher number of chances, shots and crosses used formations 1-4-2-3-1 and 1-4-3-3. Finally, most of the key passes performed in axis of the field and they were short low passes. The ability of teams to keep the possession of the ball, looking for the key pass, and final action can be a factor in the team's success. A characteristic finding of the study was the preference of groups to manifest their attack from the axis. The equal distribution of attacks between the axis and the sides of the field makes the team's offensive function more unpredictable. Also, most formations concentrate several players on the axis, making it more difficult to complete an attack from this area.

Research on international logistics supply chain management strategy based on deep reinforcement learning

Abstract The use of deep reinforcement learning algorithms for strategy formulation in supply chain management enables the nodes in the supply chain to better improve their management strategies. In this paper, a supply chain model is constructed as a starting point, and deep reinforcement learning algorithms are introduced on this basis. Firstly, the decision problem of uncertainty is handled by the reinforcement learning method of functions, and the DQN algorithm (deep neural network algorithm) is divided into two parts for iterative rules. Then the target network is established to make the iterative process more stable, to improve the convergence of the algorithm, evaluate the loss function in the training process of the network, and to determine its influence factor. Then the neural network is used to improve the iteration rule, improve the output layer, select the final action, and define the model expectation reward. Finally, the Bellman equation is fitted to the function by a deep neural network to calculate the final result. The experimental results show that by analyzing and constructing the cost of international logistics under supply chain management, the capacity utilization rate of ocean freight link is 57% The unloading link is 74% and the total capacity utilization rate is calculated as 76%. It shows that using deep reinforcement learning algorithms under international logistics supply chain management is feasible and necessary for improving the management strategy research of supply chains.

Bounded rationality, enactive problem solving, and the neuroscience of social interaction.

This article aims to show that there is an alternative way to explain human action with respect to the bottlenecks of the psychology of decision making. The empirical study of human behaviour from mid-20th century to date has mainly developed by looking at a normative model of decision making. In particular Subjective Expected Utility (SEU) decision making, which stems from the subjective expected utility theory of Savage (1954) that itself extended the analysis by Von Neumann and Morgenstern (1944). On this view, the cognitive psychology of decision making precisely reflects the conceptual structure of formal decision theory. This article shows that there is an alternative way to understand decision making by recovering Newell and Simon's account of problem solving, developed in the framework of bounded rationality, and inserting it into the more recent research program of embodied cognition. Herbert Simon emphasized the importance of problem solving and differentiated it from decision making, which he considered a phase downstream of the former. Moreover according to Simon the centre of gravity of the rationality of the action lies in the ability to adapt. And the centre of gravity of adaptation is not so much in the internal environment of the actor as in the pragmatic external environment. The behaviour adapts to external purposes and reveals those characteristics of the system that limit its adaptation. According to Simon (1981), in fact, environmental feedback is the most effective factor in modelling human actions in solving a problem. In addition, his notion of problem space signifies the possible situations to be searched in order to find that situation which corresponds to the solution. Using the language of embodied cognition, the notion of problem space is about the possible solutions that are enacted in relation to environmental affordances. The correspondence between action and the solution of a problem conceptually bypasses the analytic phase of the decision and limits the role of symbolic representation. In solving any problem, the search for the solution corresponds to acting in ways that involve recursive feedback processes leading up to the final action. From this point of view, the new term enactive problem solving summarizes this fusion between bounded and embodied cognition. That problem solving involves bounded cognition means that it is through the problem solver's enactive interaction with environmental affordances, and especially social affordances that it is possible to construct the processes required for arriving at a solution. Lastly the concept of enactive problem solving is also able to explain the mechanisms underlying the adaptive heuristics of rational ecology. Its adaptive function is effective both in practical and motor tasks as well as in abstract and symbolic ones.

IMPLEMENTATION OF EDUCATIONAL SUPERVISION TO IMPROVE TEACHER PERFORMANCE: A COMPARATIVE STUDY AT INDONESIAN AND PAKISTANI SCHOOL

Educational supervision should be conducted for various reasons one of them is to improve teachers’ perfomance as well as to improve students’ learning and quality of education. This is a comparative study of Indonesian and Pakistani school; Madrasah Ummul Quro At-Tarbawiyah Indonesia and Dar-e-Arqam School Bawalpur Punjab Pakistan. There are three research focuses, namely (1) how is the educational supervision implemented in those schools? (2) how effective is the implemenetation of educational supervision in those schools? (3) what are the problems and solutions of the implementation of the educational supervision. This study adopted a descriptive qualitative approach with interviews, observation and documentation as data collection. The results show that the director of Madrasah at Madrasah Ummul Quro At-Tarbawiyah Indonesia implemented educational supervision by using group technique through 3 activities- a monthly meetings, weekly meetings and trainings. through these forms of activity, the results are obtained in the form of positive changes from the teachers which also indirectly affect the effectiveness of the learning activities carried out so as to obtain optimal results through students’ achievements. The obstacles faced are related to the teacher's internal condition so that the alternative is by giving summons, providing motivation, and the final action is in the form of teacher transfers. Menwhile, in Dar-e-Arqam School Pakistan, head office monitoring visits implemented the supervision by monitoring visit and training session (physical and online), reports, and performance conducted monthly and in the end of semester. The head office then made the decision and guidelline for the improvement.

The One-Away Effect: The Pursuit of Mere Completion

Abstract A series of controlled studies found that consumers counter-normatively prefer something nearly complete over something complete. We call this phenomenon the “one-away effect” because we find that when consumers are, for example, one stamp away from completing a punch card loyalty program, they value the card more than a completed card. This is because their valuation of the one-away card is influenced by their anticipation of merely completing the card, which generates its own utility, apart from the card’s end-reward (a free coffee). To wit, the prospective utility of performing the final action that fulfills completion increases consumers’ valuation of the one-away card. Our findings suggest that consumers are motivated to complete goals, tasks, and sets not only to obtain their end-rewards, but also because merely completing things is intrinsically motivating and can be a goal in and of itself. We discuss the theoretical and practical implications of the one-away effect, as well as the general notion of mere completion.

TBL Myanmar: Understanding the Barriers to Sale of Dehumidifiers in Emerging Markets

Vipul Murarka, Executive Director of Tropical Biotechnology Ltd. (TBL), was perplexed after looking at the erratic sales of dehumidifiers over the past 3 years since he started in June 2016. Although the number of goods sold was growing until 2019, which gave a marginal increase in revenue earned, the top line was increasing at a decreasing rate, with 10% in the first year, 8% in the second year and 7% in the third year. He could not understand the reasons for these inconsistent sales despite offering quality solutions and having many satisfied clients in the past. Given the torrential rains throughout the year in Yangon, the capital of Myanmar, a dehumidifier was an essential product for factories, yet sales were not picking up as anticipated. Dehumidifiers as an industrial product aided many manufacturing units, as they helped in controlling moisture by extracting water from the air, thus reducing humidity and preventing the growth of mildew. There were myriad manufacturing companies in Yangon because of the availability of cheap labour, yet sales for certain months were flat, with not even one sales call funnelling to final action. These lull periods made Vipul anxious and forced him to think if he had entered the wrong industry. He even thought if the first-mover advantage turned into a disadvantage for him. The initial reason he speculated about this sales barrier was the price, as it was a high-investment product for industrial buyers. However, he also posited to himself that he might not have been able to position the advantages of the product aptly in the minds of the potential buyers. Unable to find a definitive reason for the sales barriers of the dehumidifiers, he approached his friend, Myint Than, an independent consultant and visiting faculty at an MBA institute where he taught sales and marketing strategy courses. Vipul and Myint had to find the reason for the barrier to the dehumidifier sales and eventually make a strategy to overcome these barriers for future growth before deciding the sales target for the year 2020.

An SVPWM Algorithm for a Novel Multilevel Rectifier with DC-Side Capacitor Voltage Balance

The recently proposed novel unidirectional multilevel rectifier, a three single-phase star-connected multilevel rectifier, has the characteristic of having a large number of DC-side capacitors and a complex capacitor voltage balancing control structure under conventional carrier-based phase-shift sine wave pulse-width modulation (SPWM). Hence, a space vector pulse-width modulation (SVPWM) algorithm for the novel multilevel rectifier is proposed in this paper, which can quickly balance the capacitor voltage without an additional voltage balancing control structure. Firstly, it divides the space vectors of the rectifier, then it determines the two basis voltage vectors that synthesize the output reference voltage. After that, based on the analysis of the relationship between switching states and the charge–discharge of capacitors, the final action sequences of redundant vectors are determined according to the principle of keeping the capacitor charge–discharge time consistent. Thus, the capacitor voltages can be automatically balanced without an additional voltage balancing control structure. Finally, simulation and experimental results validated the feasibility and effectiveness of the proposed SVPWM algorithm. The results also show improvements in current quality, capacitor voltage balance and the fluctuation of the neutral point voltage on the DC-link, allowing for further reduction in the overall volume and cost of the rectifier.

Quantification of Lactose in Lactose-Free and Low-Lactose Milk and Milk Products by BIOMILK 300/3000 Lac, Collaborative Study: Final Action 2020.09.

In April 2020, the BIOMILK 300 LAC method for the quantification of lactose in lactose-free and low-lactose dairy products was adopted as First Action Official Method of AnalysisSM2020.09 by AOAC INTERNATIONAL. To test the reproducibility of the method, as the last step toward the Final Action status, a collaborative study was organized by BIOLAN Microbiosensores. Fifteen collaborators within the European Union took part in this study, where nine different samples were sent as duplicate blind test portions for lactose determination. The data obtained were used to determine method repeatability and reproducibility and also to validate the new version of the biosensor. The test method is based on the direct enzymatic recognition-electrochemical detection of trace levels of lactose over a wide range of dairy samples by means of the BIOMILK 300 biosensor, in less than 5 min, and without intricate sample pretreatments. All samples resulted in a repeatability relative standard deviation (RSDr) of <8.1% and a reproducibility relative standard deviation (RSDR) of 14% maximum, meeting requirements from Standard Method Performance Requirement (SMPR®) 2018.09. The method has proved to give reproducible results for the quantification of lactose in lactose-free and low-lactose dairy products. On the basis of these results, the enzymatic amperometric biosensor method developed by BIOLAN Microbiosensores was adopted as a Final Action Official Method in July 2022.

End-to-End Autonomous Driving With Semantic Depth Cloud Mapping and Multi-Agent

Focusing on the task of point-to-point navigation for an autonomous driving vehicle, we propose a novel deep learning model trained with end-to-end and multi-task learning manners to perform both perception and control tasks simultaneously. The model is used to drive the ego vehicle safely by following a sequence of routes defined by the global planner. The perception part of the model is used to encode high-dimensional observation data provided by an RGBD camera while performing semantic segmentation, semantic depth cloud (SDC) mapping, and traffic light state and stop sign prediction. Then, the control part decodes the encoded features along with additional information provided by GPS and speedometer to predict waypoints that come with a latent feature space. Furthermore, two agents are employed to process these outputs and make a control policy that determines the level of steering, throttle, and brake as the final action. The model is evaluated on CARLA simulator with various scenarios made of normal-adversarial situations and different weathers to mimic real-world conditions. In addition, we do a comparative study with some recent models to justify the performance in multiple aspects of driving. Moreover, we also conduct an ablation study on SDC mapping and multi-agent to understand their roles and behavior. As a result, our model achieves the highest driving score even with fewer parameters and computation load. To support future studies, we share our codes at https://github.com/oskarnatan/end-to-end-driving.

Semi-Automated Enzymatic Determination of Ethanol in Beverages: Collaborative Study for RIDA&amp;#174;CUBE Ethanol

Easy and quick methods to quantify ethanol reliably in beverages are always important. In 2022, the EnzytecTM Liquid Ethanol test kit was approved as AOAC Official MethodSM 2017.07 Final Action after a collaborative study was conducted with different beverages such as kombucha, juices, and beer. During set-up of this collaborative test, small sized companies asked to include the RIDA®CUBE Ethanol/RIDA®CUBE SCAN device since it is easy to use, suitable for a few samples only and contains the identical reagents as the EnzytecTM Liquid system. It is applicable to quantify ethanol in diluted kombucha, fruit juices, and alcohol-free beer samples around 0.5% alcohol-by-volume within 12 min. The overall relative reproducibility standard deviation across a wide concentration range for kombucha, was calculated to be 6.29%. Analysis of juices and beer showed an overall higher variation with an estimated overall RSD(R) value by regression of 14.4%. The data obtained by this collaborative study show that the RIDA®CUBE Ethanol in combination with the RIDA®CUBE SCAN device is suitable to quantify ethanol from matrices representing important alcohol-free liquid food categories.

Negotiation agent based on Deep reinforcement learning for multi-agent cooperative distributed predictive control

In this work, deep neural network trained by the model-free reinforcement learning method is proposed as a negotiation agent in multi-agent distributive model predictive control. The negotiation agent is implemented in the upper layer of a hierarchical control architecture based on distributed model predictive controller with pairwise negotiation fuzzy logic. The proposed is data-driven in order to achieve the minimization of any dynamic behavior index and the specification of constraints. Specifically, the deep neural network provides consensus coefficients to address the final control action applied by each agent. The proposed is successfully applied to a hydraulic system composed of eight interconnected tanks that are very difficult to control due to non-linearities nature and high interaction between its subsystems.

Multi-level Content-aware Boundary Detection for Temporal Action Proposal Generation.

It is challenging to generate temporal action proposals from untrimmed videos. In general, boundary-based temporal action proposal generators are based on detecting temporal action boundaries, where a classifier is usually applied to evaluate the probability of each temporal action location. However, most existing approaches treat boundaries and contents separately, which neglect that the context of actions and the temporal locations complement each other, resulting in incomplete modeling of boundaries and contents. In addition, temporal boundaries are often located by exploiting either local clues or global information, without mining local temporal information and temporal-to-temporal relations sufficiently at different levels. Facing these challenges, a novel approach named multi-level content-aware boundary detection (MCBD) is proposed to generate temporal action proposals from videos, which jointly models the boundaries and contents of actions and captures multi-level (i.e., frame level and proposal level) temporal and context information. Specifically, the proposed MCBD preliminarily mines rich frame-level features to generate one-dimensional probability sequences, and further exploits temporal-to-temporal proposal-level relations to produce two-dimensional probability maps. The final temporal action proposals are obtained by a fusion of the multi-level boundary and content probabilities, achieving precise boundaries and reliable confidence of proposals. The extensive experiments on the three benchmark datasets of THUMOS14, ActivityNet v1.3 and HACS demonstrate the effectiveness of the proposed MCBD compared to state-of-the-art methods. The source code of this work can be found in https://mic.tongji.edu.cn.