CN113609992B - Analysis method of clapping motion tactical flow - Google Patents
Analysis method of clapping motion tactical flow Download PDFInfo
- Publication number
- CN113609992B CN113609992B CN202110901555.3A CN202110901555A CN113609992B CN 113609992 B CN113609992 B CN 113609992B CN 202110901555 A CN202110901555 A CN 202110901555A CN 113609992 B CN113609992 B CN 113609992B
- Authority
- CN
- China
- Prior art keywords
- tactical
- tactics
- beat
- sequence
- generator
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000004458 analytical method Methods 0.000 title abstract description 21
- 238000000034 method Methods 0.000 claims abstract description 41
- 230000008569 process Effects 0.000 abstract description 7
- 230000008859 change Effects 0.000 description 8
- 230000006870 function Effects 0.000 description 8
- 238000005516 engineering process Methods 0.000 description 7
- 230000007423 decrease Effects 0.000 description 6
- 238000005065 mining Methods 0.000 description 6
- 230000000007 visual effect Effects 0.000 description 5
- 238000010586 diagram Methods 0.000 description 4
- 238000007418 data mining Methods 0.000 description 3
- 230000006399 behavior Effects 0.000 description 2
- 230000001364 causal effect Effects 0.000 description 2
- 238000010276 construction Methods 0.000 description 2
- 230000007547 defect Effects 0.000 description 2
- 238000011282 treatment Methods 0.000 description 2
- 238000012800 visualization Methods 0.000 description 2
- 238000013473 artificial intelligence Methods 0.000 description 1
- 238000009412 basement excavation Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 238000007405 data analysis Methods 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 238000007619 statistical method Methods 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/213—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods
- G06F18/2132—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods based on discrimination criteria, e.g. discriminant analysis
Landscapes
- Engineering & Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Data Mining & Analysis (AREA)
- Theoretical Computer Science (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses an analysis method of a beat sport tactical stream, which comprises the steps of obtaining a plurality of round sequence data, wherein each round sequence data comprises a plurality of beats of data, each beat of data comprises a plurality of technical details and a beat sequence, and the plurality of technical details are used as a first tactical set; constructing a beat motion tactical set analysis system based on a generated countermeasure network, wherein the beat motion tactical set analysis system comprises a generator and a discriminator, and the generator is used for candidate tactical sets; the discriminator is used for obtaining a second tactical set, discriminating the candidate tactical set generated by the generator based on the minimum pattern length principle, returning the discrimination result to the generator, and continuously iterating the countermeasure training by the generator and the discriminator to obtain a final tactical set; converting the plurality of round sequence data into a plurality of round tactical sequence sets based upon the final tactical set; based on the multiple round tactical sequence sets, greedy strategies are adopted to obtain beat motion tactical flows. The method has better interpretation, can process multiple event sequences, and can be free from the limitation of the field of beat knowledge.
Description
Technical Field
The invention relates to the field of data analysis, in particular to an analysis method of a beat sport tactical stream.
Background
The clapping exercise refers to tennis-isolated exercise such as ping-net exercise. It is characterized in that players need to use the racket to hit balls alternately on two sides of the net. In this type of sport, the basic data unit is a shot. A series of consecutive alternating shots scored from one shot to one shot is referred to as a round. Often, each shot, the player considers a number of shot details, including the position of the players, the current position of the ball, the technique of the shot, the spin of the ball, etc. The details of the shots considered for different claps have similar parts (e.g., player position, ball position, shot technique, etc.), but not exactly the same (e.g., shuttlecocks tend to take spin into account, but take ball height into account). In each round, the player may use various tactics with variations in the details of the shots. Tactics is a short-term countermeasure strategy. For example, classical left-right mobilization tactics consist of three beats: the player of the player first hits the ball to the rightmost side of the opponent (ball position); the opponent runs to the right to get back (the position of the player); the player then uses an aggressive ball striking technique (ball striking technique) to quickly hit the ball to the far left (ball position) of the opponent, making it difficult for the opponent to return the ball. Tactics is a key point of the important analysis in the field of clapping sports.
Event sequence data is a common data type: an event is a data record of the event's body, time of occurrence, type of behavior (i.e., who did what at what time); a sequence of events is a chronological sequence of related events (as if the subject were all events in a day). An important concept in the event sequence is the feature pattern, i.e. the frequently occurring sub-sequences. The beat motion data is essentially a sequence of events: each shot may be considered an event; each round may be considered a sequence of events; the tactics frequently used by players can then be regarded as feature patterns. But beat motion is a special sequence of events: the behavior type of each event contains a record of a plurality of shot details. We refer to this sequence of events as a multivariate sequence of events.
Document J.Wang, J.Wu, A.Cao, Z.Zhou, H.Zhang, and Y.Wu.Tac-miner: visual tactic mining for multiple table tennis matches.IEEE Transactions on Visualization and Computer Graphics,27 (6): 2770-2782,2021.doi:10.1109/TVCG.2021.3074576 discloses a technique for visual mining analysis for tactics in a multi-table tennis match. It proposes a tactical definition based on domain knowledge: i.e., every three consecutive shots (i.e., my shots, opponent shots, my re-shots) in a round are considered a tactical. Based on the definition above, tac-Miner uses a traversal method to search each tactic in multiple games and statistically obtain several tactics with highest occurrence frequency. The technology has the following defects: 1. its tactical definition has limitations. The tactical definition of the technology is based on domain knowledge, not the data itself. This results in tactics that they dig into are often already known to the field specialist. At the same time, the tactical definitions proposed by this technique lead to more similar tactics, resulting in higher analysis complexity. For example, two tactics may have only one shot with different details, and three shots with the same details. Such two similar tactics may be more likely to be categorized as the same tactic in the sense of athletes and coaches. 2. The analysis method focuses on a single tactical rather than the course of change of tactical. Often times, a round of table tennis will contain multiple shots, which also results in players often applying multiple tactics to cope with opponents changing in different scenarios. The gram-to-gram relationship is one analytical focus of urgent attention in the field. For example, after a player uses a first tactic, he or she may have a different tactic choice for his or her opponent; the player will also choose different tactical treatments for different tactics of the opponent. In such complex gaming scenarios, players and coaches need to know how the opponent will choose and how my should deal with. However, the above documents only perform statistical analysis for a single tactic without going deep into the complex changes of the consequences before the tactic.
Document S.Guo, K.Xu, R.Zhao, D.Gotz, H.Zha, and N.Cao.EventThead: visual summarization and stage analysis of event sequence data.IEEE Transactions on Visualization and Computer Graphics,24 (1): 56-65,2017.doi:10.1109/TVCG.2017.2745320 discloses a visual analysis technique for pattern changes in event sequences. The technology mainly uses a method based on unsupervised learning to mine feature patterns in a sequence: it cuts a long sequence into several short sequences according to a specific time interval (for example, every 5 events), and clusters the short sequences cut by different long sequences in each time interval according to the similarity. The technology has the following defects: 1. this technique cannot handle multiple event sequences. The event of a beat sport contains a number of shot details. The unsupervised learning proposed by this technique does not effectively handle multiple event sequences. 2. The feature pattern definition of this technology has limitations based on domain knowledge. The method first cuts the sequence at specific time intervals. This is similar to the method of recognizing one tactic every three beats. 3. The result of unsupervised learning presents an interpretable problem. In artificial intelligence, interpretability is an important issue: the model resembles a black box, and the resulting results are often not known to the person who is giving the result. Nor does the present technology explain the results of its unsupervised learning. This is unacceptable to the field expert of beat sports and difficult to learn. Athletes often need to know not only how good a tactic is, but why it is so that they can understand the tactic deeply and flexibly.
Disclosure of Invention
The invention provides an analysis method of a beat sport tactic stream, which has better interpretation, can process a multi-element event sequence, and can be free from the limitation of the beat knowledge field.
A method of analyzing a beat sport tactical flow, comprising:
s1: obtaining a plurality of round sequence data, wherein each round sequence data comprises a plurality of beats of data arranged in order, each beat of data comprises a plurality of technical details and a beat order, and the plurality of technical details are used as a first tactical set;
s2: the system comprises a generator and a discriminator, wherein the generator is used for merging any two initial tactics input into a first tactic set to obtain a candidate tactic set; a discriminator for adding the candidate tactical set to the first tactical set to obtain a second tactical set, discriminating the byte lengths of the first tactical set and the second tactical set based on the minimum pattern length principle, and returning the discrimination result to the generator, wherein the generator and the discriminator continuously iterate the countermeasure training until the byte length of the N tactical set generated by the generator is not reduced, and the N tactical set is used as a final tactical set;
s3: converting the plurality of rounds of sequence data into a plurality of rounds of tactical sequence sets based upon the final tactical set, wherein each round includes a tactical sequence;
s4: based on a plurality of round tactical sequence sets, adopting a greedy strategy, forming a row of a plurality of tactics of different rounds and the same beat sequence, and merging the same tactics of the same row to obtain a beat type movement tactical stream.
The invention adopts the generation type countermeasure network, based on the analysis of the sequence data of a plurality of rounds, a plurality of tactics are accurately and reasonably obtained, the limitation of the knowledge field of the beat sports is avoided, the tactics of the invention comprise a plurality of technical details, the exhibition and analysis of the multi-element events can be provided for the clients, and the causal relationship between the tactics can be obtained through the steps S3 and S4, thereby having better interpretability.
The method for merging any two initial tactics in the first tactic set to obtain a candidate tactic set includes:
combining the plurality of attributes in the first tactical set according to different permutation and combination modes to obtain a plurality of candidate tactical sets, wherein the plurality of attributes are the same type of attributes of two initial tactics with the largest use times in the same round.
For each iteration training, the generator can combine two tactics arbitrarily selected from the tactics set obtained by the previous iteration training, and add the combined result into the tactics set obtained by the previous iteration training through the discriminator to obtain a new tactics set.
The combined initial tactics are deleted from the second tactics set to obtain a third tactics set before the second tactics set is input to the generator, and the third tactics set is input to the generator if the byte length of the third tactics set is less than the byte length of the first tactics set to delete the redundant tactics.
The distinguishing the byte length of the first tactical set and the second tactical set comprises:
the second tactical set is input to the generator if the byte length of the second tactical set is less than the byte length of the first tactical set, and returned to the first tactical set if the byte length of the second tactical set is greater than the byte length of the first tactical set.
The byte length is composed of the number of tactics in the final tactical collection, the number of tactical uses, and the number of remaining technical details from the plurality of rounds of sequence data resulting from removal of technical details in the final tactical collection.
The byte length function L (·) is:
where P represents the number of tactics in the tactical collection, u (P i S) represents the ith tactic p in the tactical collection i The total number of occurrences in the set of round sequences, sv (s i P) represents describing an ith pass sequence S in a plurality of pass sequence data using a tactical collection P i The number of shot details which cannot be described is remained, and alpha, beta and gamma are control parameters.
According to the minimum description length principle, when the value of the byte length function L is minimum, the tactical collection P can optimally describe the original round sequence dataset, i.e., the tactical collection at this time can best summarize the tactical characteristics of the athlete. To sum up the tactics that are most valuable for analysis, we have chosen three properties that are most relevant to tactics for defining a byte length function: 1. the tactical number describes the diversity of tactical uses by athletes during play: the actual tactics number of a player is generally 10-20; 2. the number of tactical uses illustrates the analytical value of tactical: if the tactics are used too little, the tactics are not common in practice and lack of analysis value; if the tactics are used too many times, the tactics themselves contain less batting details and lack analytical value (e.g. a tactics containing only a single batting will necessarily appear many times, but one batting cannot reflect the tactical thinking of the player); 3. the number of shot details remaining describes the tactical descriptive ability of the round data: the remaining undescribeable shot details are too many, indicating that the resultant tactics do not describe the original round data well; otherwise, the description tactics is well descriptive of the original round data. In an actual iterative update process, the three terms contained in the byte length function remain dynamically balanced, i.e., as several terms decrease, others must increase. For example: 1. as the number of tactics decreases and the number of tactics maneuvers decreases, there is a potential for more shot details not being described; 2. as the number of tactics decreases, so too does the remaining undescribed details of the shots, then each tactic must necessarily be used more times; 3. when the number of tactical uses decreases, and the details of the remaining shots decrease, it is stated that the overall tactical is almost the same as the original round sequence, and the number of tactics must increase if each original round corresponds to a new tactical. We use three control parameters to maintain this dynamic balance: we generally fix α to 1 and then adjust the values of β and γ to balance tactical diversity, analytical value, descriptive capacity.
Forming a column of multiple tactics in different rounds and in the same beat order, comprising:
the beat order is the order of each beat in each round, and a plurality of tactics of the same beat order in different rounds are formed into a column, wherein if the beat orders included in the two tactics are not overlapped, the latter tactic is in a column with respect to the former tactic.
Compared with the prior art, the invention has the beneficial effects that:
the invention adopts the generation type countermeasure network, based on the analysis of the sequence data of a plurality of rounds, a plurality of tactics are accurately and reasonably obtained, the limitation of the knowledge field of the beat sports is avoided, the tactics of the invention comprise a plurality of technical details, the exhibition and analysis of the multi-element events can be provided for the clients, and the causal relationship between the tactics can be obtained through the steps S3 and S4, thereby having better interpretability.
Drawings
FIG. 1 is a tactical definition diagram provided by an embodiment;
FIG. 2 is a flow chart of a method for analyzing a beat sport tactical flow according to an embodiment;
fig. 3 is a diagram of a candidate tactical generation method provided in an embodiment.
Detailed Description
We define a tactic as a frequently occurring "non-overlapping, incompletely contiguous short subsequence". As shown in fig. 1, tactical t is in two rounds s 1 Sum s 2 All are used, s 1 Sum s 2 Is the sequence data of two rounds (each containing four shot events, labeled e 1 ～e 4 The method comprises the steps of carrying out a first treatment on the surface of the Each batting event contains two attributes, the batting technique and the ball position, respectively), t is the tactical dug in the two rounds. In this definition, the subsequence is the core, and appears as tactics as well as a battingAnd the sequence is part of the original round sequence data and the order remains consistent. For example, a round containing 5 shots may be 1 st, 2 nd, or 3 rd, 4 th, 5 th shots. The non-overlapping is embodied in that the two tactics do not contain the same shot in the original round. For example, a round containing 5 shots, it is impossible that beats 1, 2, 3 are a tactical, and beats 2, 3, 4 are a tactical. This is because an athlete may only perform one tactic at each shot, and may not perform two different tactics simultaneously. An incomplete manifestation is that each shot in a tactical does not fully cover all of the shot details. For example, classical left-right mobilization tactics involve 3 beats: the player of the player first hits the ball to the rightmost side of the opponent (ball position); the opponent runs to the right to get back (the position of the player); the player then uses an aggressive ball striking technique (ball striking technique) to quickly hit the ball to the far left (ball position) of the opponent, making it difficult for the opponent to return the ball. It is important to have several shot details per shot, and other shot details may vary. The shots that appear continuously as tactical must be connected in the original round sequence data without missing any shots in the middle. This is because in a beat-like sport, the athlete's thinking is often linear and does not skip a beat of thinking tactics. For example, beats 1, 2, and 4 of one round, while also subsequences, lack of beat 3 cannot be considered tactical. Short shots, which are represented as tactics, typically do not exceed 5 beats. This is because the beat sport changes rapidly, and a tactic of more than 5 beats requires not only thorough execution by the player, but even perfect coordination by the opponent, which is difficult to occur in a real game.
2. The technology provides a data mining algorithm flow for mining the tactics and constructing tactical variant processes.
Such tactical definitions are data driven, unlike any previous definitions, so mining such tactics requires a specialized data mining algorithm. Meanwhile, in order to study the tactical course of change, the data mining algorithm also needs to have a corresponding mechanism. As shown in fig. 2, the algorithm flow includes four phases. Initial stage (a): we first obtain a beat motion event sequence dataset (4 rounds, 4 rounds containing 4, 8, 7, 8 shots in turn, each shot taking into account 2 shot details). Tactical excavation phase (B): we excavate several tactics from the original dataset (3 tactics X, Y, Z are excavated in the figure). Sequence conversion stage (C): we mark each tactical use in the original sequence, and then for each round, convert the sequence of shots into a tactical sequence in the order in which the tactical uses. Diagram construction stage (D): according to the tactical sequence, we will merge similar tactics and construct a directed acyclic graph.
The following expansion explains the tactical mining algorithm and tactical change procedure flow diagram construction algorithm.
1. Tactical mining algorithms. This algorithm is mainly used to mine all tactics from the original round sequence dataset. The technique herein employs a concept similar to that of creating an countermeasure network, i.e., we use a generator to continually create new candidate tactics, while using a discriminant to determine if a tactic is well descriptive of the player's strategy. Through multiple iterations, a final stable tactical collection is finally obtained.
A generator: the generator first initializes a collection of tactics, each of which is a single shot detail (i.e., a shot containing only one particular shot detail, such as using the loop technique, catching a ball in a backhand area, etc.). Then, in each iteration, the generator selects two tactics in the tactical collection to attempt to merge. The generator preferably selects the two tactics that are used the most frequently at the same time in the same round. Fig. 3 illustrates a method of merging new tactics, i.e., merging two selected tactics together in different alignments. It should be noted that if some alignment results in a collision of values (e.g., cp 5 ) Such alignment may be discarded.
A discriminator: the discriminant uses the principle of minimum description length to decide whether a tactical should be added to a tactical collection. The core of the arbiter is a function L (S, P), where S is the original round sequence set, P is the tactical set, and the L () function is the byte length needed when describing the round sequence set S based on the tactical set P. When a generator generates a candidate tactic, the arbiter adds the candidate tactic to tactic collection P and compares the change in byte length before and after the addition. If the byte length becomes shorter after addition, then it is advantageous to describe the original round sequence to illustrate the candidate tactics. And when a candidate tactic is added to the tactic collection, the arbiter further attempts to eliminate two tactics from the tactic collection for constructing the candidate tactic to avoid redundant tactics. If deleting a tactic can make the byte length shorter, then that tactic is redundant and should be deleted. Therefore, the core of the arbiter lies in the definition of the function. We calculate three pieces of information in this function that are of most interest in the field of beat motion in visual analysis: first, the number of tactics; second, the number of tactical uses; third, coverage of tactics. This can be achieved by:
where P represents the number of tactics in the tactical collection, u (P i S) represents the ith tactic p in the tactical collection i The total number of occurrences in the set of round sequences, sv (s i P) represents describing an ith pass sequence S in a plurality of pass sequence data using a tactical collection P i The number of shot details which cannot be described is remained, and alpha, beta and gamma are control parameters.
2. Tactical change procedure builds algorithms. The algorithm is mainly used for constructing a tactical change process based on the round sequence data set and the tactical collection. We first label each tactical use in the original round sequence, and then for each round, convert the sequence of shots into a tactical sequence in the order in which the tactics were used. After deriving several tactical sequences, we construct a tactical transformation process by aligning and merging tactical two steps. We first used a greedy strategy to align tactics therein. The greedy principle comprises two points: first, each tactic is aligned as far forward as possible to ensure a compact, easy analysis of the overall change process. Secondly, if there is no overlap between the beats of two tactics, a list should be put in place for the tactics that occur later, and cannot be aligned together. Finally, if there is some identical tactic in the different tactic sequences and aligned, we merge the tactics to simplify the course of tactic change.
Claims (7)
1. A method of analyzing a beat sport tactical flow, comprising:
s1: obtaining a plurality of round sequence data, wherein each round sequence data comprises a plurality of beats of data arranged in order, each beat of data comprises a plurality of technical details and a beat order, and the plurality of technical details are used as a first tactical set;
s2: the system comprises a generator and a discriminator, wherein the generator is used for merging any two initial tactics input into a first tactic set to obtain a candidate tactic set; a discriminator for adding the candidate tactical set to the first tactical set to obtain a second tactical set, discriminating the byte lengths of the first tactical set and the second tactical set based on the minimum pattern length principle, and returning the discrimination result to the generator, wherein the generator and the discriminator continuously iterate the countermeasure training until the byte length of the N tactical set generated by the generator is not reduced, and the N tactical set is used as a final tactical set;
s3: converting the plurality of rounds of sequence data into a plurality of rounds of tactical sequence sets based upon the final tactical set, wherein each round includes a tactical sequence;
s4: based on a plurality of round tactical sequence sets, adopting a greedy strategy to form a row of a plurality of tactics of different rounds and the same beat sequence, and combining the same tactics of the same row to obtain a beat type movement tactical stream;
the byte length function L () is:
where P represents the number of tactics in the tactical collection, u (P i S) represents warIth tactical p in tactical collection i The total number of occurrences in the set of round sequences, sv (s i P) represents describing an ith pass sequence S in a plurality of pass sequence data using a tactical collection P i The number of shot details which cannot be described is remained, and alpha, beta and gamma are control parameters.
2. The method of claim 1, wherein the step of merging any two initial tactics of the input first tactics set to obtain a candidate tactics set comprises:
combining the plurality of attributes in the first tactical set according to different permutation and combination modes to obtain a plurality of candidate tactical sets, wherein the plurality of attributes are the same type of attributes of two initial tactics with the largest use times in the same round.
3. The method of claim 1, wherein the generator combines two tactics arbitrarily selected from the tactics set obtained in the previous iteration, and adds the combined result to the tactics set obtained in the previous iteration to obtain a new tactics set.
4. The method of claim 1, wherein the combined initial tactics are deleted from the second tactical collection to obtain a third tactical collection before the second tactical collection is input to the generator, and the third tactical collection is input to the generator to delete the redundant tactics if the byte length of the third tactical collection is less than the byte length of the first tactical collection.
5. The method of claim 1, wherein the step of determining the byte length of the first tactical set and the second tactical set comprises:
the second tactical set is input to the generator if the byte length of the second tactical set is less than the byte length of the first tactical set, and returned to the first tactical set if the byte length of the second tactical set is greater than the byte length of the first tactical set.
6. The method of claim 1, wherein the byte length is comprised of a number of tactics in the final tactical collection, a number of tactical uses, and a number of remaining technical details from the plurality of rounds of sequence data removed from the technical details in the final tactical collection.
7. The method of claim 1, wherein forming a column of a plurality of tactics of different rounds and the same beat order, comprises:
the beat sequence is the sequence of each beat in each round, and a plurality of tactics of the same beat sequence in different rounds are formed into a column, wherein if the beat sequences included in two tactics are not overlapped, the latter tactic is in a column relative to the former tactic.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110901555.3A CN113609992B (en) | 2021-08-06 | 2021-08-06 | Analysis method of clapping motion tactical flow |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110901555.3A CN113609992B (en) | 2021-08-06 | 2021-08-06 | Analysis method of clapping motion tactical flow |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113609992A CN113609992A (en) | 2021-11-05 |
| CN113609992B true CN113609992B (en) | 2023-11-14 |
Family
ID=78339652
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110901555.3A Active CN113609992B (en) | 2021-08-06 | 2021-08-06 | Analysis method of clapping motion tactical flow |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113609992B (en) |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108463271A (en) * | 2015-08-28 | 2018-08-28 | 伊虎智动有限责任公司 | System and method for motor skill analysis and technical ability enhancing and prompt |
| CN110727826A (en) * | 2019-09-30 | 2020-01-24 | 浙江大学 | Visual analysis method for tactical excavation of table tennis |
| CN112363849A (en) * | 2020-10-23 | 2021-02-12 | 中国电子科技集团公司第三十研究所 | Lightweight service interaction protocol method in tactical environment |
| CN113158390A (en) * | 2021-04-29 | 2021-07-23 | 北京邮电大学 | Network attack traffic generation method for generating countermeasure network based on auxiliary classification |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US11620919B2 (en) * | 2019-02-11 | 2023-04-04 | Sierra Nevada Corporation | Live virtual constructive gateway systems and methods |
| US11113535B2 (en) * | 2019-11-08 | 2021-09-07 | Second Spectrum, Inc. | Determining tactical relevance and similarity of video sequences |
-
2021
- 2021-08-06 CN CN202110901555.3A patent/CN113609992B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108463271A (en) * | 2015-08-28 | 2018-08-28 | 伊虎智动有限责任公司 | System and method for motor skill analysis and technical ability enhancing and prompt |
| CN110727826A (en) * | 2019-09-30 | 2020-01-24 | 浙江大学 | Visual analysis method for tactical excavation of table tennis |
| CN112363849A (en) * | 2020-10-23 | 2021-02-12 | 中国电子科技集团公司第三十研究所 | Lightweight service interaction protocol method in tactical environment |
| CN113158390A (en) * | 2021-04-29 | 2021-07-23 | 北京邮电大学 | Network attack traffic generation method for generating countermeasure network based on auxiliary classification |
Non-Patent Citations (1)
| Title |
|---|
| 图卷积网络的行为识别技术在空手道技战术分析中的应用;郭建平 等;《湖南师范大学自然科学学报》;第44卷(第1期);第81-87页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113609992A (en) | 2021-11-05 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Parmar et al. | Learning to score olympic events | |
| Katona et al. | Time to die: Death prediction in dota 2 using deep learning | |
| Wu et al. | TacticFlow: Visual analytics of ever-changing tactics in racket sports | |
| Wang et al. | Exploring the long short-term dependencies to infer shot influence in badminton matches | |
| Ghosh et al. | Sports analytics review: Artificial intelligence applications, emerging technologies, and algorithmic perspective | |
| Hughes | How to develop a notation system | |
| Shen | A machine learning approach to predict the result of League of Legends | |
| Wang et al. | How is the stroke? inferring shot influence in badminton matches via long short-term dependencies | |
| Xenopoulos et al. | Graph neural networks to predict sports outcomes | |
| Browne et al. | Discovering patterns of play in netball with network motifs and association rules | |
| Rahimian et al. | Beyond action valuation: A deep reinforcement learning framework for optimizing player decisions in soccer | |
| Liu et al. | On efficient reinforcement learning for full-length game of starcraft ii | |
| Bamunif | Sports information & discussion forum using artificial intelligence techniques: A new approach | |
| Cao et al. | Team-builder: Toward more effective lineup selection in soccer | |
| Ding et al. | Deep reinforcement learning in a racket sport for player evaluation with technical and tactical contexts | |
| CN113609992B (en) | Analysis method of clapping motion tactical flow | |
| Elmiligi et al. | Predicting the outcome of soccer matches using machine learning and statistical analysis | |
| Riza et al. | Determining strategies on playing badminton using the Knuth-Morris-Pratt algorithm | |
| US11790654B2 (en) | Derived competition | |
| Vis et al. | Tennis patterns: Player, match and beyond | |
| Beheshtian-Ardakani et al. | CMPN: Modeling and analysis of soccer teams using Complex Multiplex Passing Network | |
| Liang | Research on prediction of the game winner based on artificial intelligence methods | |
| Ravi et al. | Using Game Theory to maximize the chance of victory in two-player sports | |
| Ban et al. | BadmintonDB: A Badminton Dataset for Player-specific Match Analysis and Prediction | |
| Sari et al. | Modeling the Badminton Stroke Pattern Through the Sequential Pattern Discovery Using Equivalent Classes (SPADE) Algorithm |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| CB03 | Change of inventor or designer information |
Inventor after: Wu Yingcai Inventor after: Wu Jiang Inventor after: Zhang Hui Inventor before: Wu Yingcai Inventor before: Wu Jiang |
|
| CB03 | Change of inventor or designer information | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |