An electronic sports game review result generation method and device, electronic equipment and storage medium

Abstract

The application provides an electronic sports game session review result generation method and device, electronic equipment and a storage medium, comprising: obtaining at least one key segment of a target electronic sports game session; performing data analysis on each key segment to construct a segment feature set; retrieving and matching a corresponding rule knowledge subset from a pre-established rule knowledge base to determine a review conclusion label corresponding to the key segment; determining a reason explanation and an evidence chain based on the review conclusion label, the segment feature set and the rule knowledge subset; determining an optimization suggestion according to the review conclusion label, the reason explanation and the rule knowledge subset; and binding the review conclusion label, the reason explanation, the evidence chain and the optimization suggestion with time axis information of the key segment to determine a structured review result. The application can automatically convert the key segment in the electronic sports game session into a structured review result containing a conclusion, a reason, evidence and a suggestion, which significantly improves the efficiency compared with manual review.

Description

Technical Field

[0001] This application relates to the field of e-sports data processing technology, and more specifically, to a method, apparatus, electronic device, and storage medium for generating e-sports game replay results. Background Technology

[0002] Esports matches are characterized by their fast-paced nature, rapid changes in character status, frequent resource competition, and complex teamwork. A single crucial team fight, lane rotation, resource exchange, skill usage error, or positioning mistake can often have a cascading effect on the subsequent game. Therefore, post-match analysis requires more than just reviewing the video; it also necessitates explaining the rationality of actions at key moments, the reasons for mistakes, and how to correct them.

[0003] Currently, common methods for reviewing esports match replays include manual review of recordings, viewing match data, in-game replays, highlight clips, and coach commentary. While these methods offer some reference, they still have significant limitations: First, manual review heavily relies on coach experience, leading to inconsistent judgments of the same segment among different personnel; second, match data and highlight clips typically only reflect the outcome, failing to explain the process that led to it; third, review conclusions lack a stable link to the video timeline, requiring users to repeatedly drag the video to find corresponding evidence; fourth, many comments remain subjective descriptions such as "played poorly" or "made poor decisions," lacking actionable improvement suggestions; fifth, review results are difficult to standardize, limiting subsequent retrieval, statistical analysis, training plan development, and ability profile generation.

[0004] With the development of artificial intelligence, video analytics, and data processing technologies, some tools have begun to attempt automated analysis of game videos. However, most solutions focus on image recognition, event detection, or data visualization, lacking methods for organizing replay results for training scenarios. Even if a system can identify kills, deaths, team fights, or resource changes, it does not equate to being able to output replay conclusions that can be used for training. Results with higher training value should simultaneously include conclusion labels, causal chains, evidence nodes, and optimization suggestions, and be linked to the video timeline.

[0005] Therefore, existing technologies need a solution that can uniformly process key moments in esports events, character states, situational context, resource information, and rule knowledge, and automatically generate structured replay results. Summary of the Invention

[0006] This invention aims to overcome the shortcomings of the prior art and provide a method, device, electronic device, and storage medium for generating e-sports game review results. This addresses the technical problems in the prior art, such as the reliance on human experience in generating conclusions for e-sports game reviews, inconsistent output structures, opaque cause analysis, insufficient feasibility of suggestions, and difficulty in binding review results to the video timeline.

[0007] In a first aspect, the present invention provides a method for generating replay results of e-sports matches, the method comprising: Obtain at least one key segment of the target esports match; Data parsing is performed on each of the key segments to construct a segment feature set; Based on the game type of the target e-sports match and the segment type of the key segment, retrieve and match the corresponding rule knowledge subset from the pre-established rule knowledge base, match the segment feature set with the rule knowledge subset, and determine the replay conclusion label corresponding to the key segment; Based on the debriefing conclusion labels, the fragment feature set, and the rule knowledge subset, the explanation of the cause and the chain of evidence are determined. Based on the review conclusion labels, the explanations of reasons, and the subset of rule knowledge, optimization suggestions are determined; The debriefing conclusion labels, the explanations of causes, the chain of evidence, and the optimization suggestions are bound to the timeline information of the key segments to determine the structured debriefing results.

[0008] Preferably, acquiring at least one key segment of the target esports match includes: Acquire video data of the target esports match; Extract the key segments from the video data; The key segments include the start time, the end time, and the segment type; The types of clips include kill clips, death clips, team fight clips, resource contest clips, lane switching clips, vision control clips, skill release clips, or abnormal operation clips.

[0009] Preferably, the construction of the fragment feature set includes: Extract event features, character status features, situation features, and resource features to form the fragment feature set; The event characteristics include event type, event time, participating players, kills, deaths, assists, skill usage, and target contention information; The character's status characteristics include health, energy, level, equipment, location, skill cooldown status, and economic disadvantage; The situation features include the number of players on both sides, map location, minion wave status, vision coverage, and differences in team resources; The resource characteristics include the status of Dragon, Herald, Baron, turrets, and jungle resources.

[0010] Preferably, determining the debriefing conclusion label corresponding to the key segment includes: Calculate a confidence score for each candidate label based on the fragment feature set and the rule knowledge subset; Candidate labels whose confidence scores reach a preset threshold are determined as the review conclusion labels; When there are multiple candidate labels with confidence scores reaching a preset threshold, the review conclusion labels are output in descending order of confidence scores. The debriefing conclusion labels include correct operation, operational error, decision-making problem, improper resource management, positioning problem, vision problem, pacing problem, skill release problem, communication and coordination problem, or risk control problem.

[0011] Preferably, determining the cause explanation and evidence chain based on the debriefing conclusion labels, the fragment feature set, and the rule knowledge subset includes: Features related to the debriefing conclusion label are selected from the fragment feature set, and judgment rules related to the debriefing conclusion label are selected from the rule knowledge subset; The selected feature changes, rule conditions, and final effects are combined to form the explanation of the reasons. Evidence nodes are extracted from the fragment feature set and multiple evidence nodes are arranged in chronological order to form the evidence chain; Each of the aforementioned evidence nodes includes evidence time, evidence object, evidence type, and evidence description.

[0012] Preferably, determining optimization suggestions based on the review conclusion labels, the explanations of causes, and the subset of rule knowledge includes: Based on the debriefing conclusion tags, the explanation of reasons, and the suggested template rules in the suggested template subset, candidate suggested templates are retrieved from the suggested template library; Based on the specific characteristics of the key segments and the tactical rules in the rule knowledge subset, the role name, time point, resource name and status condition in the candidate suggestion template are filled in to determine action-oriented natural language suggestions; The optimization suggestions include suggestion types and suggestion content. Suggestion types include positioning adjustments, skill release order adjustments, resource exchange strategies, retreat timing, vision control, communication and coordination, rhythm shifts, or risk avoidance.

[0013] Preferably, after determining the structured debriefing results, the method further includes: The structured review results are stored in a non-transitory computer-readable storage medium; In response to a user's click on any conclusion on the debriefing page, the system automatically jumps to the corresponding video segment in the video data based on the timeline anchor point in the structured debriefing results, and simultaneously displays the evidence node and the optimization suggestions.

[0014] Secondly, the present invention provides an apparatus for generating e-sports game replay results, comprising: The key segment acquisition module is used to acquire at least one key segment of the target e-sports match; The fragment feature set construction module is used to parse data for each of the key fragments and construct a fragment feature set. The debriefing conclusion label determination module is used to retrieve and match the corresponding rule knowledge subset from a pre-established rule knowledge base based on the game type of the target e-sports game and the segment type of the key segment, match the segment feature set with the rule knowledge subset, and determine the debriefing conclusion label corresponding to the key segment. The cause explanation and evidence chain determination module is used to determine the cause explanation and evidence chain based on the review conclusion label, the fragment feature set, and the rule knowledge subset. The optimization suggestion determination module is used to determine optimization suggestions based on the review conclusion labels, the explanation of reasons, and the rule knowledge subset. The structured debriefing result determination module is used to bind the debriefing conclusion labels, the explanations of causes, the evidence chain, and the optimization suggestions with the timeline information of the key segments to determine the structured debriefing result.

[0015] Thirdly, the present invention provides a readable medium including executable instructions, which, when executed by a processor of an electronic device, cause the electronic device to perform any of the methods described in the first aspect.

[0016] Fourthly, the present invention provides an electronic device including a processor and a memory storing execution instructions, wherein when the processor executes the execution instructions stored in the memory, the processor performs the method as described in any of the first aspects.

[0017] This invention provides a method, apparatus, electronic device, and storage medium for generating esports game replay results. It integrates key segments, segment events, player status, situation information, resource information, and rule knowledge, ensuring that replay conclusions no longer rely solely on human experience or single data displays, thus improving the stability and consistency of conclusion generation. Employing a structured output method of "conclusion label—reason explanation—optimization suggestions," the replay results simultaneously answer "what the problem was, why the problem was, and how to improve next," enhancing training feasibility. By binding replay results to video segments through evidence nodes and timeline anchors, users can directly review corresponding evidence, reducing the time spent manually searching for segments. It supports standardized storage and retrieval of replay results, facilitating the subsequent creation of historical replay records, ability profiles, trend analyses, and training plans. It can be deployed on cloud platforms, desktop software, mobile applications, or local training systems, exhibiting good engineering implementation adaptability.

[0018] The further effects of the aforementioned non-conventional preferred method will be explained below in conjunction with specific embodiments. Attached Figure Description

[0019] To more clearly illustrate the embodiments of the present invention or the existing technical solutions, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0020] Figure 1 This is a schematic diagram of a method for generating replay results of e-sports games according to an embodiment of the present invention; Figure 2 This is a schematic diagram of another method for generating e-sports game replay results according to an embodiment of the present invention; Figure 3 This is a schematic diagram of the composition of an e-sports game replay result generation device according to an embodiment of the present invention; Figure 4 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention. Detailed Implementation

[0021] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below in conjunction with specific embodiments and corresponding drawings. Obviously, the described embodiments are only a part of the embodiments of this invention, and not all of them. Based on the embodiments of this invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this invention.

[0022] See Figure 1 The image shows a specific embodiment of a method for generating esports match replay results provided by the present invention. In this embodiment, the method for generating esports match replay results includes:

[0023] Step 101: Obtain at least one key segment of the target esports match; Specifically, the system first acquires video data from the target esports match. This video data can originate from local uploads, platform recordings, game replay files, or cloud video addresses. Then, key segments are extracted from the video data. These key segments can be manually annotated or automatically located by event detection modules, video frame change detection modules, or kill event detection modules based on screen changes, kill notifications, audio peaks, resource refresh notifications, character status changes, or game logs. Each key segment includes a start time, end time, and segment type. Segment types include kill segments, death segments, team fight segments, resource contention segments, lane switching segments, vision control segments, skill usage segments, or abnormal operation segments. For example, in a match, the system automatically locates a resource contention segment with a start time of 08:35 and an end time of 08:52; its segment type is "resource contention segment." The start and end times are used to precisely link the replay conclusions with the video timeline later.

[0024] Step 102: Perform data parsing on each key segment and construct a segment feature set; Furthermore, for each key segment, data analysis is performed to extract multi-dimensional feature information, forming a segment feature set. This feature set includes at least event features, character status features, situation features, and resource features. Event features record key behaviors occurring within the segment, including event type, event time, participating players, kills, deaths, assists, skill usage information, and target contention information. Character status features reflect the real-time status of participating characters, including health, energy (or mana), level, equipment list, location coordinates, skill cooldown status, and economic difference. Situation features describe the overall situation of both sides, including the number of players on both sides, map distribution, minion wave status, vision coverage, and team resource differences (such as differences in the number of turrets and dragon souls). Resource features focus on the status of key resources on the map, including the spawn status or ownership of dragons, heralds, barons, turrets, and jungle resources.

[0025] In the practical example, for the resource contention segment obtained in step 101, the system extends the preset time window before the start time T1 and after the end time T2 to obtain the segment context interval, and then extracts the following features: the friendly jungler's Smite is on cooldown, the enemy mid laner has been missing for more than 8 seconds, the target has 15 seconds remaining to respawn, the friendly river vision is empty, the friendly team has 4 players, and the enemy team has 5 players. The system divides the above features into static features (character, equipment, level, lineup, target type) and dynamic features (position change, health change, skill release, movement trajectory, number of players change, target ownership change), and encodes them into structured key-value data, such as recording "Friendly jungler's Smite status: on cooldown", "Enemy mid laner's disappearance time: 8 seconds", "Target respawn countdown: 15 seconds", "Friendly river vision: none", and "Number of players: 4:5". These features provide a data foundation for subsequent judgment on whether the resource contention is reasonable.

[0026] Step 103: Based on the game type and key segment type of the target e-sports match, retrieve and match the corresponding rule knowledge subset from the pre-established rule knowledge base, match the segment feature set with the rule knowledge subset, and determine the replay conclusion label corresponding to the key segment. Specifically, firstly, based on the game type and the type of key segment in the current match (such as a resource contention segment), the system retrieves rule knowledge relevant to the scenario from a pre-established rule knowledge base, forming a subset of rule knowledge. This rule knowledge base includes basic game rules, character skill rules, resource refresh rules, team composition counter rules, common tactical rules, training evaluation rules, risk assessment rules, and suggested template rules. For example, for a resource contention segment, the system prioritizes resource contention rules (such as unit difference rules, vision rules, and punishment status rules); for a team fight segment, the system prioritizes team fight rules (such as positioning rules, focus fire rules, and key skill rules); and for a death segment, the system prioritizes death cause analysis rules (such as vision loss rules, skill cooldown rules, and retreat timing rules).

[0027] The system then matches the feature set of the segment with the subset of rule knowledge, calculating a confidence score for each candidate debriefing conclusion label. The confidence score is weighted by event matching score, state anomaly score, resource impact score, rule matching score, and context impact score. For example, in the resource contention segment mentioned above, the system deducts points based on the difference in the number of players (4 friendly, 5 enemy), lack of vision, penalty cooldown, and the imminent target refresh (high risk), resulting in a comprehensive calculation: a confidence score of 0.86 for the "Inappropriate Resource Management" label, 0.72 for the "Positioning Problem" label, and 0.15 for the "Correct Operation" label. When the preset threshold is 0.7, the system determines "Inappropriate Resource Management" as the primary debriefing conclusion label. If multiple labels simultaneously reach the threshold, they are output in descending order of confidence score. Debriefing conclusion labels include correct operation, operational error, decision-making problem, inappropriate resource management, positioning problem, vision problem, rhythm problem, skill release problem, communication and coordination problem, or risk control problem.

[0028] Step 104: Based on the debriefing conclusion labels, fragment feature sets, and rule knowledge subsets, determine the explanation of causes and the chain of evidence; Specifically, after determining the conclusion labels of the review, further explanations of the causes and chains of evidence are generated.

[0029] First, features related to the current replay conclusion label are selected from the fragment feature set. For example, if the replay conclusion label is "risk control issue", the system selects features such as vision coverage status, enemy disappearance information, character position coordinates, and the feasibility of retreat routes. At the same time, judgment rules related to this label are selected from the rule knowledge subset, such as "when the river vision is missing and the enemy jungler has disappeared for more than 5 seconds, continuing to push forward constitutes a high-risk behavior".

[0030] Secondly, the selected feature changes, rule conditions, and final impacts are combined to form a cause explanation. The cause explanation is organized according to the structure of "situation background - key behavior - direct consequence - training problem". For example, in a death segment, the cause explanation generated by the system is: "The enemy jungler disappeared in the upper half of the map (situation background), our team lacked vision in the river and the target character continued to push forward to farm minions (key behavior), resulting in being surrounded and killed by two enemy players (direct consequence). This death is a risk control problem caused by the superposition of vision loss and route selection (training problem)."

[0031] Then, evidence nodes are extracted from the fragment feature set. Each evidence node includes four elements: evidence time (specific time point, such as T-8 seconds, T-5 seconds, T-3 seconds, time T), evidence object (such as enemy jungler, friendly character, river vision), evidence type (such as state change, event occurrence, position movement), and evidence description (such as "enemy jungler disappeared in the top half of the map", "friendly river vision is missing", "target character continues to push forward to farm minions", "killed by two enemies"). The system arranges multiple evidence nodes in chronological order to form an evidence chain. For example, for the aforementioned death sequence, the chain of evidence is as follows: {time:T-8,object:enemy jungler,type:state change,desc:disappeared in the top half of the map}; {time:T-5,object:river vision,type:vision loss,desc:no vision in the friendly river}; {time:T-3,object:target character,type:behavior,desc:continued to push forward and farm minions}; {time:T,object:target character,type:event,desc:killed by two enemies}. This chain of evidence allows the review conclusions to be traced back to specific time points and state changes, making it easier for users to understand the basis of the conclusions.

[0032] Step 105: Based on the review conclusion labels, explanations of causes, and subsets of rule knowledge, determine optimization suggestions; Furthermore, based on the debriefing conclusion tags and causal explanations, and combined with the suggestion template rules in the rule knowledge subset, matching candidate suggestion templates are retrieved from the suggestion template library. The suggestion template library pre-stores standardized suggestion sentences for different problem types. For example, for the "risk control problem," the retrieved suggestion template is: "When [condition 1] and [condition 2], execute [action]; the training objective is [objective]." Then, based on the specific characteristics of the key segments and the tactical rules in the rule knowledge subset, the system fills in the role name, time point, resource name, and status conditions in the candidate suggestion templates, generating action-oriented natural language suggestions. The system fills in condition 1 as "enemy jungler disappears," condition 2 as "river vision is missing," the action as "prioritize retreating to a safe line, do not continue pushing forward to farm minions; if you must deal with the minion wave, wait for the support or jungler to replenish vision before advancing," and the training objective as "reduce the number of deaths when pushing forward without vision," ultimately generating optimization suggestions. For example, regarding "skill release issues," the system generates the following suggestion: "Do not release key control skills before the enemy uses their movement skills; force out movement skills or wait for teammates to follow up with control skills before releasing them." Regarding "improper resource management," the system generates the following suggestion: "Complete vision control 30 seconds before the target spawns; if your team is one player short and lacks key skills, choose to exchange resources or clear the lane instead of forcibly engaging in a team fight." The optimization suggestions include suggestion types and suggestion content. Suggestion types cover positioning adjustments, skill release order adjustments, resource exchange strategies, retreat timing, vision control, communication and coordination, rhythm shifting, and risk avoidance.

[0033] Step 106: Bind the review conclusion labels, explanations of reasons, evidence chains, and optimization suggestions to the timeline information of key segments to determine the structured review results.

[0034] Specifically, the system associates and binds the debriefing conclusion labels, explanations of reasons, evidence chains, and optimization suggestions generated in steps 103 to 105 with timeline information such as the start and end times of the key segment, forming a structured debriefing result entity. For example, for the above resource contention segment, the system generates the following structured data: the segment number is S20260401_001, the start time is 08:35, the end time is 08:52, the replay conclusion is tagged as improper resource handling, the reason is explained as "one less player on our side, lack of vision in the river, key skills on cooldown, still choosing to enter the resource area head-on", the evidence node list is [{time:08:38, object:enemy, type:number of players, desc:enemy has more players}, {time:08:42, object:vision, type:missing, desc:our side lacks vision}, {time:08:48, object:our side, type:control, desc:controlled first}], the optimization suggestion is "vision should be set up in advance before the target refreshes, when the number of players is insufficient, priority should be given to exchanging resources or clearing lanes, and forced team fights should be avoided", the confidence score is 0.86, the display priority is high, and the associated video address is the match video URL. The structured review results can be parsed by the front-end interface and used for subsequent display, retrieval, and linked playback.

[0035] As can be seen from the above technical solution, the beneficial effect of this embodiment is that it automatically converts key segments in e-sports games into structured review results containing conclusions, reasons, evidence, and suggestions, which significantly improves efficiency, consistency, and training guidance value compared to manual review.

[0036] Figure 1 The embodiments shown are merely basic examples of the method of the present invention. Other preferred embodiments of the method can be obtained by making certain optimizations and extensions based on them.

[0037] like Figure 2 The image shows another specific embodiment of the method for generating esports game replay results according to the present invention. This embodiment further describes the method based on the foregoing embodiments, and includes the following steps:

[0038] Step 201: Store the structured review results to a non-transitory computer-readable storage medium; Specifically, after generating the structured debriefing result in step 106 of the previous embodiment, the system persistently stores the result in a non-transitory computer-readable storage medium. The storage medium can be a hard drive, solid-state drive, flash memory, memory card, or cloud storage. The storage format of the structured debriefing result can be JSON, XML, or a database record. Each result includes fields such as segment number, start time, end time, debriefing conclusion label, reason explanation, evidence node list (each node includes evidence time, evidence object, evidence type, and evidence description), optimization suggestions, confidence score, display priority, and associated video address. For example, the structured result of the resource contention segment mentioned above is stored in JSON format, where the segment number is "S20260401_001", the confidence score is 0.86, and the display priority is "high". After storage, this data can be used for retrieving historical review records, constructing player ability profiles (statistically analyzing the frequency of the same conclusion label appearing in different games for the same player), trend analysis of multiple games (judging the improvement of players on a certain type of problem), and formulating training plans (arranging special training for high-frequency problems), thus achieving the reusability of review results.

[0039] Step 202: In response to the user's click on any conclusion on the review page, automatically jump to the corresponding video segment in the video data based on the timeline anchor point in the structured review results, and simultaneously display evidence nodes and optimization suggestions.

[0040] Furthermore, in the front-end replay interface, the system displays a video player, a timeline component, and structured replay result cards. Each replay result card displays a replay conclusion label (e.g., "Improper resource handling"), a key reason explanation (e.g., "One less player on our side, lack of vision in the river"), and a summary of optimization suggestions (e.g., "Vision should be deployed before the target refreshes"). When a user clicks on any conclusion card, the system automatically controls the video player to jump to that time position based on the pre-bound timeline anchor point in the structured replay result (i.e., the start time of the key segment 08:35 or the specific time point corresponding to the evidence node, such as 08:38, 08:42, 08:48), and highlights the corresponding node on the timeline. At the same time, the complete reason explanation, evidence chain (users can expand to view detailed information for each evidence node, such as "08:38 Enemy has more players", "08:42 Our side lacks vision", "08:48 We were controlled first"), and complete optimization suggestions are displayed in the sidebar or floating window of the interface. Users can further click on a specific evidence node in the evidence chain (for example, clicking "08:42 Our Vision is Missing"), and the system will jump back to the precise time point (08:42) corresponding to that evidence node, enabling multi-level linked review and helping users accurately understand every step of the problem.

[0041] As can be seen from the above technical solutions, the beneficial effects of this embodiment are: through storage and interactive linkage, the review results can not only be recorded and reused, but also closely cooperate with the video timeline, greatly reducing the time users spend manually searching for segments and improving review efficiency and learning experience.

[0042] This invention also provides an apparatus for generating replay results of esports matches. See also: Figure 3 The image shows a specific embodiment of an esports match replay result generation device provided by the present invention. This embodiment of the device is used to execute... Figures 1-2 The physical apparatus of the method. Its technical solution is essentially the same as the above embodiments, and the corresponding descriptions in the above embodiments also apply to this embodiment. The apparatus includes:

[0043] The key segment acquisition module 301 is configured to acquire at least one key segment of the target e-sports game; The fragment feature set construction module 302 is configured to perform data parsing on each key fragment and construct a fragment feature set; The review conclusion label determination module 303 is configured to retrieve and match the corresponding rule knowledge subset from a pre-established rule knowledge base based on the game type and the segment type of the key segment of the target e-sports game, match the segment feature set with the rule knowledge subset, and determine the review conclusion label corresponding to the key segment. The cause explanation and evidence chain determination module 304 is configured to determine the cause explanation and evidence chain based on the review conclusion labels, fragment feature set and rule knowledge subset; The optimization suggestion determination module 305 is configured to determine optimization suggestions based on the review conclusion labels, cause descriptions, and rule knowledge subsets. The structured debriefing result determination module 306 is configured to bind the debriefing conclusion labels, cause explanations, evidence chains, and optimization suggestions with the timeline information of key segments to determine the structured debriefing result.

[0044] Figure 4 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention. At the hardware level, the electronic device includes a processor, and optionally also includes an internal bus, a network interface, and a memory. The memory may include main memory, such as high-speed random-access memory (RAM), or it may also include non-volatile memory, such as at least one disk storage device. Of course, the electronic device may also include other hardware required for other services.

[0045] The processor, network interface, and memory can be interconnected via an internal bus, which can be an ISA (Industry Standard Architecture) bus, a PCI (Peripheral Component Interconnect) bus, or an EISA (Extended Industry Standard Architecture) bus, etc. Buses can be categorized as address buses, data buses, and other types. For ease of representation, Figure 4 The symbol is represented by a single double-headed arrow, but this does not mean that there is only one bus or one type of bus.

[0046] Memory is used to store instructions for execution. Specifically, instructions for execution are computer programs that can be executed. Memory can include main memory and non-volatile memory, and it provides the processor with execution instructions and data.

[0047] In one possible implementation, the processor reads the corresponding execution instructions from non-volatile memory into main memory and then executes them. Alternatively, it may obtain the corresponding execution instructions from other devices to form an e-sports game replay result generation device at the logical level. The processor executes the execution instructions stored in the memory to implement the e-sports game replay result generation method provided in any embodiment of the present invention through the executed execution instructions.

[0048] The above is as described in the present invention. Figure 3 The method executed by the e-sports game replay result generation device provided in the illustrated embodiment can be applied to a processor, or implemented by a processor. The processor may be an integrated circuit chip with signal processing capabilities. During implementation, each step of the above method can be completed by integrated logic circuits in the processor's hardware or by instructions in software form. The processor can be a general-purpose processor, including a Central Processing Unit (CPU), a Network Processor (NP), etc.; it can also be a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), a Field-Programmable Gate Array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components. It can implement or execute the methods, steps, and logic block diagrams disclosed in the embodiments of this invention. The general-purpose processor can be a microprocessor or any conventional processor.

[0049] The steps of the method disclosed in the embodiments of this invention can be directly manifested as being executed by a hardware decoding processor, or executed by a combination of hardware and software modules in the decoding processor. The software modules can reside in random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, registers, or other mature storage media in the art. This storage medium is located in memory, and the processor reads information from the memory and, in conjunction with its hardware, completes the steps of the above method.

[0050] This invention also proposes a readable medium storing execution instructions. When these instructions are executed by a processor of an electronic device, the device can perform an e-sports game replay result generation method provided in any embodiment of this invention, specifically for executing, as... Figure 1 , Figure 2 The method shown.

[0051] The electronic devices in the foregoing embodiments may be computers.

[0052] Those skilled in the art will understand that embodiments of the present invention can be provided as methods or computer program products. Therefore, the present invention can be implemented in a completely hardware embodiment, a completely software embodiment, or a combination of software and hardware.

[0053] The various embodiments in this invention are described in a progressive manner. Similar or identical parts between embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. In particular, the apparatus embodiments are basically similar to the method embodiments, so the description is relatively simple; relevant parts can be referred to the descriptions of the method embodiments.

[0054] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0055] The above are merely embodiments of the present invention and are not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principle of the present invention should be included within the scope of the claims of the present invention.

Claims

1. A method for generating replay results of esports matches, characterized in that, The method includes: Obtain at least one key segment of the target esports match; Data parsing is performed on each of the key segments to construct a segment feature set; Based on the game type of the target e-sports match and the segment type of the key segment, retrieve and match the corresponding rule knowledge subset from the pre-established rule knowledge base, match the segment feature set with the rule knowledge subset, and determine the replay conclusion label corresponding to the key segment; Based on the debriefing conclusion labels, the fragment feature set, and the rule knowledge subset, the explanation of the cause and the chain of evidence are determined. Based on the review conclusion labels, the explanations of reasons, and the subset of rule knowledge, optimization suggestions are determined; The debriefing conclusion labels, the explanations of causes, the chain of evidence, and the optimization suggestions are bound to the timeline information of the key segments to determine the structured debriefing results.

2. The method according to claim 1, characterized in that, The acquisition of at least one key segment of the target esports match includes: Acquire video data of the target esports match; Extract the key segments from the video data; The key segments include the start time, the end time, and the segment type; The types of clips include kill clips, death clips, team fight clips, resource contest clips, lane switching clips, vision control clips, skill release clips, or abnormal operation clips.

3. The method according to claim 1, characterized in that, The constructed fragment feature set includes: Extract event features, character status features, situation features, and resource features to form the fragment feature set; The event characteristics include event type, event time, participating players, kills, deaths, assists, skill usage, and target contention information; The character's status characteristics include health, energy, level, equipment, location, skill cooldown status, and economic disadvantage; The situation features include the number of players on both sides, map location, minion wave status, vision coverage, and differences in team resources; The resource characteristics include the status of Dragon, Herald, Baron, turrets, and jungle resources.

4. The method according to claim 1, characterized in that, The debriefing conclusion tags corresponding to the key segments include: Calculate a confidence score for each candidate label based on the fragment feature set and the rule knowledge subset; Candidate labels whose confidence scores reach a preset threshold are determined as the review conclusion labels; When there are multiple candidate labels with confidence scores reaching a preset threshold, the review conclusion labels are output in descending order of confidence scores. The debriefing conclusion labels include correct operation, operational error, decision-making problem, improper resource management, positioning problem, vision problem, pacing problem, skill release problem, communication and coordination problem, or risk control problem.

5. The method according to claim 1, characterized in that, The determination of the cause explanation and evidence chain based on the debriefing conclusion labels, the fragment feature set, and the rule knowledge subset includes: Features related to the debriefing conclusion label are selected from the fragment feature set, and judgment rules related to the debriefing conclusion label are selected from the rule knowledge subset; The selected feature changes, rule conditions, and final effects are combined to form the explanation of the reasons. Evidence nodes are extracted from the fragment feature set and multiple evidence nodes are arranged in chronological order to form the evidence chain; Each of the aforementioned evidence nodes includes evidence time, evidence object, evidence type, and evidence description.

6. The method according to claim 1, characterized in that, The step of determining optimization suggestions based on the review conclusion labels, the explanations of causes, and the subset of rule knowledge includes: Based on the debriefing conclusion tags, the explanation of reasons, and the suggested template rules in the suggested template subset, candidate suggested templates are retrieved from the suggested template library; Based on the specific characteristics of the key segments and the tactical rules in the rule knowledge subset, the role name, time point, resource name and state condition in the candidate suggestion template are filled in to determine action-oriented natural language suggestions; The optimization suggestions include suggestion types and suggestion content. The suggestion types include positioning adjustments, skill release order adjustments, resource exchange strategies, retreat timing, vision control, communication and coordination, rhythm shifts, or risk avoidance.

7. The method according to any one of claims 1 to 6, characterized in that, After determining the structured debriefing results, the process also includes: The structured review results are stored in a non-transitory computer-readable storage medium; In response to a user's click on any conclusion on the debriefing page, the system automatically jumps to the corresponding video segment in the video data based on the timeline anchor point in the structured debriefing results, and simultaneously displays the evidence node and the optimization suggestions.

8. A device for generating replay results of e-sports matches, characterized in that, include: The key segment acquisition module is used to acquire at least one key segment of the target e-sports match; The fragment feature set construction module is used to parse data for each of the key fragments and construct a fragment feature set. The debriefing conclusion label determination module is used to retrieve and match the corresponding rule knowledge subset from a pre-established rule knowledge base based on the game type of the target e-sports game and the segment type of the key segment, match the segment feature set with the rule knowledge subset, and determine the debriefing conclusion label corresponding to the key segment. The cause explanation and evidence chain determination module is used to determine the cause explanation and evidence chain based on the review conclusion label, the fragment feature set, and the rule knowledge subset. The optimization suggestion determination module is used to determine optimization suggestions based on the review conclusion labels, the explanation of reasons, and the rule knowledge subset. The structured debriefing result determination module is used to bind the debriefing conclusion labels, the explanations of causes, the evidence chain, and the optimization suggestions with the timeline information of the key segments to determine the structured debriefing result.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium includes a stored program, wherein the program, when executed, performs the method of any one of claims 1 to 7.

10. An electronic device, characterized in that, The electronic device includes: processor; Memory used to store the processor's executable instructions; The processor is configured to read the executable instructions from the memory and execute the instructions to implement the method described in any one of claims 1 to 7.

Images