Learning video playing method, device and electronic equipment

Abstract

The application provides a learning video playing method and device and electronic equipment, and relates to the technical field of online education. The method comprises the following steps: acquiring a learning video and dividing the learning video into multiple knowledge intervals arranged according to a preset learning order according to teaching units; receiving a playing instruction of the learning video, analyzing and acquiring a historical exit playing position corresponding to the playing instruction; collecting and determining the thermal characteristics of candidate knowledge intervals; combining the thermal characteristics to determine the learning degree of the candidate knowledge intervals; wherein the candidate knowledge intervals comprise a preceding knowledge interval and a current knowledge interval; determining the preceding knowledge interval or the current knowledge interval as a target knowledge interval according to the learning degree of the candidate knowledge intervals, and playing the target knowledge interval, so that the mastery degree of each knowledge interval in the video is evenly distributed, the effectiveness of short-time learning is improved, and the technical effect of eliminating understanding faults caused by resuming playing from the middle of the knowledge interval is achieved.

Description

Technical Field

[0001] This invention relates to the field of online education technology, and in particular to a method, apparatus, and electronic device for playing learning videos. Background Technology

[0002] The existing learning video playback solutions simply record the video timestamp when the user exits to achieve simple breakpoint resume playback. This can easily lead to problems such as users with weak prior knowledge continuing to play the current knowledge point, causing a learning gap, or users who have mastered the prior knowledge still having to repeat the review, wasting learning time. Overall, the level of intelligence is low, and it is difficult to match the user's actual learning situation for adaptive playback.

[0003] Therefore, in view of the shortcomings of related technical solutions, the present invention provides a method for playing learning videos. Summary of the Invention

[0004] Therefore, it is necessary to provide a learning video playback method, device, and electronic device to address the aforementioned technical problems.

[0005] On the one hand, a method for playing learning videos is provided. The method includes: acquiring learning videos and dividing them into multiple knowledge intervals arranged according to a preset learning order based on teaching units; receiving user playback instructions for the learning videos, parsing and acquiring the historical exit playback position corresponding to the playback instruction; determining candidate knowledge intervals from the multiple knowledge intervals based on the historical exit playback position, wherein the candidate knowledge intervals include previous knowledge intervals and the current knowledge interval, the current knowledge interval representing the knowledge interval corresponding to the historical exit playback position, and the previous knowledge interval representing the knowledge interval arranged before the current knowledge interval; collecting and determining the heat features of the candidate knowledge intervals based on preset heat feature rules; determining the user's learning level of the candidate knowledge intervals by combining the heat features of the candidate knowledge intervals; and determining the previous knowledge interval or the current knowledge interval as the target knowledge interval based on the learning level of the candidate knowledge intervals, and playing the target knowledge interval.

[0006] Optionally, the learning level includes mastery; based on the learning level of the candidate knowledge interval, the preceding knowledge interval or the current knowledge interval is determined as the target knowledge interval, including: in response to detecting that the mastery of a preceding knowledge interval is less than or equal to a mastery threshold, the first preceding knowledge interval whose mastery is less than or equal to the mastery threshold according to the preset learning order is taken as the target knowledge interval; in response to the mastery of a preceding knowledge interval being greater than the mastery threshold, the current knowledge interval is taken as the target knowledge interval.

[0007] Optionally, playing the target knowledge interval includes: in response to the preceding knowledge interval being the target knowledge interval, starting playback from the beginning position of the preceding knowledge interval; or, in response to the current knowledge interval being the target knowledge interval, determining the interval type of the current knowledge interval, determining the playback position of the current knowledge interval based on the interval type, and starting playback from the playback position.

[0008] Optionally, the interval type includes theoretical explanation and example explanation; the playback position of the current knowledge interval is determined based on the interval type, including: in response to the interval type of the current knowledge interval being theoretical explanation, playback starts from the beginning position of the current knowledge interval; in response to the interval type of the current knowledge interval being example explanation, the example segment corresponding to the historical exit playback position is determined, and playback starts from the beginning position of the example segment.

[0009] Optionally, the learning video is acquired and divided into multiple knowledge segments arranged according to a preset learning order by teaching unit. This includes: parsing the learning video to obtain the text content, audio explanations, and page transition timestamps in each display interface of the learning video; dividing the learning video into multiple initial knowledge segments based on the page transition timestamps; and adjusting the multiple initial knowledge segments based on the audio explanations and text content, combined with the segment duration rules, to determine the multiple knowledge segments of the learning video, with the multiple knowledge segments arranged according to a preset learning order.

[0010] Optionally, the method further includes: adjusting multiple initial knowledge intervals in conjunction with interval duration rules, including: merging multiple initial knowledge intervals into a candidate knowledge interval in response to the initial knowledge intervals meeting a first condition, wherein the first condition includes the duration of the initial knowledge interval being less than the minimum preset interval duration, and / or multiple initial knowledge intervals containing the same knowledge point; splitting the initial knowledge interval into multiple candidate knowledge intervals in response to the initial knowledge intervals meeting a second condition, wherein the second condition includes the duration of the initial knowledge interval being greater than the maximum preset interval duration, and / or the initial knowledge interval containing multiple independent knowledge points.

[0011] Optionally, the method further includes: in response to the user ending the viewing of the learning video, updating and storing the knowledge interval identifier and the offset within the knowledge interval corresponding to the historical exit position; collecting the user's learning playback behavior data, wherein the learning playback behavior data includes at least one of the following: effective viewing time, number of revisits, number of exits, number of entries, number of effective pauses, and number of complete views of the knowledge interval in the learning video; and updating and storing the learning level of the knowledge interval based on the learning playback behavior data.

[0012] Optionally, in response to learning playback behavior data including the effective viewing time of knowledge segments in learning videos, user learning playback behavior data is collected, including: obtaining the user's learning duration factor for knowledge segments; wherein, the learning duration factor includes at least one of continuous dwell time, replay time, jump-in dwell time, and jump-out time; and using preset weights to perform weighted fusion of the learning duration factor to obtain the effective viewing time of the knowledge segment.

[0013] On the other hand, a learning video playback device is provided, the device comprising: an interactive module for receiving a user's playback command for the learning video; and a computing module including a processor and a memory; the computing module is connected to the interactive module for implementing any of the above-mentioned learning video playback methods.

[0014] In another aspect, this application also provides an electronic device, including: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of any of the above-mentioned learning video playback methods; and a display device connected to the learning video playback device for displaying and playing learning videos.

[0015] The aforementioned learning video playback method, device, and electronic device include: acquiring a learning video and dividing it into multiple knowledge segments arranged according to a preset learning order based on teaching units; receiving a user's playback command for the learning video, parsing and acquiring the historical exit position corresponding to the playback command; determining candidate knowledge segments from the multiple knowledge segments based on the historical exit position, wherein the candidate knowledge segments include a preceding knowledge segment and a current knowledge segment, the current knowledge segment representing the knowledge segment corresponding to the historical exit position, and the preceding knowledge segment representing the knowledge segment arranged before the current knowledge segment; collecting and determining the heat features of the candidate knowledge segments based on preset heat feature rules; determining the user's learning level of the candidate knowledge segments based on the heat features of the candidate knowledge segments; determining the preceding knowledge segment or the current knowledge segment as the target knowledge segment based on the learning level of the candidate knowledge segments, and playing the target knowledge segment; thus, achieving a balanced distribution of mastery of each knowledge segment in the video, supporting fragmented learning at the knowledge segment level, improving the effectiveness of short-term learning, ensuring complete contextual connection, and eliminating comprehension gaps caused by continuing playback from the middle of a knowledge segment. Attached Figure Description

[0016] To more clearly illustrate the embodiments of this application, the accompanying drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 This is a schematic diagram of the architecture of a learning video playback system in one embodiment; Figure 2 This is a flowchart illustrating a video playback method in one embodiment; Figure 3 This is a schematic diagram of PPT-driven knowledge interval segmentation in a video playback learning method in one embodiment. Figure 4 This is a schematic diagram of the logical breakpoint detection and resume playback process of a video playback method in one embodiment; Figure 5 This is a structural block diagram of a learning video playback device in one embodiment; Figure 6 This is a diagram of the internal structure of an electronic device in one embodiment. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions in the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0019] It should be understood that, in the description of this application, unless the context explicitly requires it, words such as "including" or "comprising" throughout the specification should be interpreted as including rather than exclusive or exhaustive; that is, meaning "including but not limited to".

[0020] It should also be understood that the terms "first," "second," etc., are used for descriptive purposes only and should not be construed as indicating or implying relative importance. Furthermore, in the description of this application, unless otherwise stated, "a plurality of" means two or more.

[0021] It should be noted that the terms "S1," "S2," etc., are used only for descriptive purposes and do not specifically refer to the order or sequence, nor are they intended to limit this application. They are merely for the convenience of describing the method of this application and should not be construed as indicating the sequential order of the steps. Furthermore, the technical solutions of the various embodiments can be combined with each other, but this must be based on the ability of those skilled in the art to implement them. When the combination of technical solutions is contradictory or impossible to implement, it should be considered that such a combination of technical solutions does not exist and is not within the scope of protection claimed in this application.

[0022] In one embodiment, such as Figure 1As shown, a learning video playback system is provided, which includes the following architecture: video parsing layer, behavior acquisition layer, intelligent breakpoint layer, and playback resumption layer.

[0023] Specifically, the video parsing layer includes PPT (PowerPoint) page timeline parsing, knowledge point logical interval segmentation, completion model construction, and video duration statistics.

[0024] Specifically, the behavior collection layer includes collecting user dwell time records, jump / jump time points, pause / replay behavior, and heat map weight calculation.

[0025] Specifically, the intelligent breakpoint layer includes a logical starting point localization engine, a mastery assessment model, breakpoint state persistence, and personalized recommendation logic.

[0026] Specifically, the resume playback recovery layer includes precise location of logical breakpoints, fragmented learning paths, weak mastery prompt pop-ups, and progress visualization output.

[0027] In one embodiment, such as Figure 2 As shown, a method for playing learning videos is provided, including the following steps: S101: Obtain learning videos and divide them into multiple knowledge sections arranged according to a preset learning order, based on teaching units.

[0028] Here, a teaching unit is a segment of a teaching activity divided according to knowledge themes and cognitive logic.

[0029] Here, the pre-defined learning sequence is a sequential dependency relationship that must be followed between teaching units, predefined by the course designer or domain knowledge.

[0030] Specifically, the segmentation methods may include visual content parsing, auditory content parsing, and / or metadata alignment.

[0031] Here, a knowledge interval is a video segment in a learning video that is occupied by a single, continuous, and indivisible teaching activity.

[0032] The knowledge intervals can be divided before each playback, or the same learning video can be pre-divided into intervals and stored, and then the knowledge intervals can be directly called later.

[0033] Specifically, if a video segment can stand alone as a complete learning object, and does not contain another sub-segment with a complete teaching loop, then it is a knowledge interval. For example, a segment that explains the definition, derivation of principles, and description of characteristics around a knowledge point can be divided into a knowledge interval. Alternatively, one or more example problem segments around the same knowledge point can be divided into a knowledge interval, where the example problem segments present the example, analyze the thought process, demonstrate the solution process, and verify the answer.

[0034] S102: Receive the user's playback command for the learning video, parse it and obtain the historical exit position corresponding to the playback command.

[0035] Here, the play command is the operation command issued by the user to watch the learning video.

[0036] The playback command can be for the first playback of a learning video, for continuing playback of a learning video, or for selecting a chapter of a learning video to play.

[0037] Here, the historical exit position is the precise point in the video content where the user last ended watching, as recorded by the system.

[0038] Among them, the historical exit position is persistently stored using a binary knowledge interval ID and the offset within the interval to achieve cross-device consistency: the logical interval ID does not depend on the video player and is completely consistent across mobile phones / PCs / tablets and semantic stability: even if the video is re-edited (such as inserting an opening title), as long as the knowledge interval content remains unchanged, the historical exit position is still valid.

[0039] Thus, based on the structured storage scheme that reconstructs breakpoint records from a single physical timestamp into knowledge interval identifiers and offsets within the interval, the learning video playback system can establish a precise semantic mapping between timestamps and knowledge points within the video, thereby gaining the ability to perceive the knowledge context of breakpoint locations.

[0040] Specifically, the system continuously monitors user interaction events on the player interface. When a user performs actions such as opening a course or continuing to learn, the player frontend generates a playback command. The playback command is then parsed to extract the video identifier and user account identifier bound to the video. The system determines whether the user has any unfinished learning progress for the video. If a valid record is found, the system parses out the knowledge interval identifier and the offset within the knowledge interval. If this is the user's first time playing the video or the record has expired, the knowledge interval identifier and the offset within the knowledge interval will be empty, and playback will start from the beginning of the first knowledge interval of the video according to the default logic.

[0041] S103: Based on the historical exit position, determine the candidate knowledge interval from multiple knowledge intervals. The candidate knowledge interval includes the previous knowledge interval and the current knowledge interval. The current knowledge interval represents the knowledge interval corresponding to the historical exit position, and the previous knowledge interval represents the knowledge interval arranged before the current knowledge interval.

[0042] Here, the preceding knowledge interval is a knowledge interval arranged in a preset learning order before the knowledge interval corresponding to the historical exit playback position, and has a direct or indirect dependency relationship with that interval.

[0043] Here, the current knowledge interval refers to the knowledge interval corresponding to the historical exit position, that is, the knowledge interval that the user was learning when the last viewing was interrupted.

[0044] In this process, the corresponding knowledge interval is identified by a knowledge interval identifier, which is a unique number or mark for the knowledge interval. For example, "Chapter 3 - Conceptual Foundations - Conceptual Beginning".

[0045] Each knowledge zone has a unique knowledge zone identifier, zone_id, and one knowledge zone identifier, zone_id, uniquely corresponds to one knowledge zone.

[0046] In one embodiment, the top N intervals that the current knowledge interval directly depends on are determined through the knowledge graph and used as candidate knowledge intervals, instead of mechanically including all historical intervals. This pruning strategy is usually suitable for long videos or long courses, balancing the comprehensiveness of the evaluation with the system's response speed.

[0047] S104: Based on preset thermal feature rules, collect and determine the thermal features of candidate knowledge intervals.

[0048] Here, the heatmap is a comprehensive representation obtained by weighting multi-dimensional behavioral data such as the user's continuous dwell time, active replay time, time spent jumping in and out, number of times they pause to think within a given area. The heatmap directly reflects the user's attention and engagement with the knowledge area. High-heat areas usually mean that the content is important or difficult to understand, and users watch or stay in them repeatedly; low-heat areas may indicate that the content has been mastered or skipped.

[0049] Specifically, preset weights are configured for each multidimensional behavioral data point, and the thermal characteristics of the knowledge interval are obtained through weighted calculation.

[0050] S105: Combine the heat map characteristics of candidate knowledge intervals to determine the user's learning level of candidate knowledge intervals.

[0051] The level of learning can include mastery and interval type.

[0052] S106: Based on the learning level of the candidate knowledge intervals, determine the preceding knowledge interval or the current knowledge interval as the target knowledge interval, and play the target knowledge interval.

[0053] The location of the target knowledge interval to be played can include the beginning of the first unmastered preceding interval (offset of 0); a specific location within the currently exited interval (which may be the start of the interval or a backtracking point near the exit point); or the historical exit location itself (precise continuation playback).

[0054] Specifically, the system screens knowledge segments with high priority and low mastery. When a weak knowledge segment is identified, the starting playback position is set to the starting point of the segment with an offset of 0, forcing pre-existing reinforcement playback. When both the current knowledge segment and the preceding knowledge segments are mastered, the offset of the previous exit playback position is set as the starting playback position, and playback continues normally. When the current exit segment is not mastered but all preceding segments are mastered, playback starts from the logical starting point or example of the current exit segment.

[0055] In one embodiment, the user's learning progress across all courses is visualized in the form of a heatmap of knowledge interval completion, with green (≥80%) / orange (60-80%) / red (<60%) three-color coding intuitively indicating the mastery status.

[0056] In the above implementation, this application determines whether it is necessary to start learning from the previous knowledge interval before the previous playback position based on the mastery level and interval type of each knowledge interval, and determines the corresponding starting playback position, so as to achieve a balanced distribution of mastery level of each knowledge interval in the video, support fragmented learning at the knowledge interval granularity, improve the effectiveness of short-term learning, ensure complete context connection, and eliminate the understanding gap caused by continuing playback from the middle of the knowledge interval.

[0057] In some specific implementations, the learning level includes mastery; based on the learning level of the candidate knowledge interval, the prior knowledge interval or the current knowledge interval is determined as the target knowledge interval, including: In response to the detection that there is a prior knowledge interval with a mastery level less than or equal to the mastery level threshold, the first prior knowledge interval with a mastery level less than or equal to the mastery level threshold is taken as the target knowledge interval according to the preset learning order. If the mastery of the prior knowledge interval is greater than the mastery threshold, the current knowledge interval is taken as the target knowledge interval.

[0058] Here, mastery level is an indicator of a user's understanding and proficiency with content in a certain knowledge area.

[0059] In one embodiment, mastery depends not only on the degree of mastery but also on the time decay of mastery. For example, if the mastery was 90% a month ago, it may now be downgraded to "needs consolidation" due to the forgetting curve, and the decayed mastery can be calculated using the Ebbinghaus forgetting curve model.

[0060] For example, if the first interval has been completed, the second interval has been skipped, the third interval has been completed, and only one segment of the fourth interval has been watched and has not met the standard, if the completion rate of the second interval is 0 and below the threshold, the next playback will prioritize positioning to the starting point of the second interval; only when the second interval has been manually confirmed, met the test standard, or been explicitly marked as skipped by the user will the playback continue to the fourth interval.

[0061] In one embodiment, the system can dynamically adjust the mastery threshold based on the dependence strength of knowledge intervals: if the current interval has an extremely high dependence coefficient on a certain preceding interval (it will be impossible to learn without mastering the preceding interval), the judgment threshold of that preceding interval can be temporarily increased (e.g., from 60 points to 80 points), and a stricter backoff strategy can be implemented. Conversely, if the dependence is weak, the threshold can be lowered to reduce unnecessary backoff and improve learning efficiency.

[0062] Thus, based on the hierarchical judgment mechanism of target knowledge intervals driven by mastery thresholds, the learning video playback system can prioritize locking the weakest preceding intervals with knowledge gaps in the learning path as relearning targets when performing positioning, forcibly supplementing basic cognition before progressing to the current content; only when all preceding dependencies have been met will the current interval be used as the direct continuation target; thereby effectively avoiding the understanding gap caused by ignoring the mastery status of preceding knowledge during continuation playback, and building a rigorous cognitive guarantee link through bottom-up achievement verification, thereby significantly improving the logical rigor of the learning process and the semantic rationality of the continuation starting point.

[0063] In some specific implementations, the target knowledge range to be played includes: Responding to the premise knowledge interval being the target knowledge interval, playback begins from the starting position of the premise knowledge interval; or, In response to the current knowledge interval being the target knowledge interval, the interval type of the current knowledge interval is determined, and the playback position of the current knowledge interval is determined based on the interval type, and playback starts from the playback position.

[0064] In this way, by forcibly rebuilding the foundation from scratch based on the prior knowledge interval and positioning the current knowledge interval according to the fractal, while avoiding the contextual understanding gap caused by mid-stream replay, the effective learning input already completed by the user is maximized. This achieves a leap from physical time point reproduction to teaching semantic adaptation in the replay strategy, significantly improving learning coherence and cognitive efficiency.

[0065] In some specific implementations, the interval type includes theoretical explanations and example explanations; the playback position of the current knowledge interval is determined based on the interval type, including: If the current knowledge segment is classified as theoretical explanation, playback will begin from the beginning of the current knowledge segment. In response to the current knowledge interval type, the example problem is explained, and the example problem segment corresponding to the historical exit position is determined. Playback starts from the beginning position of the example problem segment.

[0066] Theoretical explanations may include concept explanations and summaries.

[0067] When explaining interval types using examples, a knowledge interval can include one or more example segments.

[0068] For example, theoretical explanations can continue from the starting point of the knowledge section; example problem explanations can continue from the starting anchor point of the example problem. The starting anchor point of the example problem can be the page turn time of a new PPT page, or the time when the question number appears in the same PPT page, the animation is triggered, the teacher marks it, or the example problem is semantically recognized by the subtitles.

[0069] Thus, for theoretical explanation sections with strong logical dependencies and progressive narrative characteristics, the playback starting point is forcibly reset to the beginning position of the section to reconstruct a complete cognitive derivation context, fundamentally eliminating the conceptual understanding gap caused by resuming playback from the middle; for example explanation sections with independent closed-loop teaching structures, the system automatically identifies the single example to which the user's exit position belongs and accurately jumps to the starting anchor point of that example, allowing users to focus directly on the unfinished problem-solving task without repeating the explanation process of the completed example; achieving a refined balance between teaching continuity and learning efficiency, significantly reducing the repeated viewing cost and cognitive switching burden caused by fragmented playback.

[0070] In some specific implementations, the learning videos are acquired and divided into multiple knowledge segments arranged according to a preset learning order, including: Analyze the learning video to obtain the text content, audio explanations, and page transition timestamps within each display interface of the learning video; Based on the page switching timestamp, the learning videos are divided into multiple initial knowledge intervals; Based on the audio and text content of the explanations, and combined with the interval duration rules, multiple initial knowledge intervals are adjusted to determine multiple knowledge intervals for the learning videos. These multiple knowledge intervals are arranged according to a preset learning order.

[0071] Here, the display interface is a single-frame static teaching screen rendered and output in real time by the front-end player during the learning video playback process. It can include visual teaching pages such as courseware slides, code editors, whiteboards, or graphic handouts.

[0072] Here, the text content is the visible text information extracted from the display interface, such as PPT titles, body text, code snippets, chart annotations, etc., which can be obtained from video frames through OCR (Optical Character Recognition) technology.

[0073] Here, the audio explanation refers to the voice content of the teacher or explainer in the video, which is usually transcribed into text using ASR (Automatic Speech Recognition) technology.

[0074] Here, page transition timestamps are records of the moments when the displayed interface in a learning video undergoes significant changes. For example, the global video time corresponding to moments such as turning a PowerPoint slide, switching code windows, or flipping through handouts.

[0075] Specifically, the input video is fully analyzed, and the following tasks are performed simultaneously: keyframes are extracted at a certain frequency (e.g., 1 frame per second or sampling when a sudden change in the scene is detected), text recognition is performed on each frame to obtain the text content in the display interface and the corresponding video timestamp; the video audio track is extracted, speech recognition is performed, and a sequence of explanatory text with timestamps is obtained; by comparing the similarity of adjacent keyframes (e.g., histogram differences), the timestamps of drastic changes in the scene are detected and recorded as page switching timestamps; the video is cut into continuous segments, i.e., candidate knowledge intervals, with the page switching timestamps as boundaries, and their start time, end time, and corresponding page screenshots are recorded.

[0076] For example, assuming the display interface is a PowerPoint presentation, the system backend extracts time data from the instructor's PowerPoint slides. Using slide page turn timestamps, on-page question anchors, subtitle semantic boundaries, and audio pauses as candidate boundaries, the video is segmented into several semantically complete knowledge intervals. The correspondence between PowerPoint slides and knowledge intervals is not fixed: one knowledge interval can cover multiple PowerPoint slides, and one PowerPoint slide can be divided into multiple knowledge intervals when it contains multiple examples or sub-questions.

[0077] Thus, the video structuring segmentation scheme based on multimodal parsing and duration semantic fusion enables the learning video playback system to automatically transform unstructured teaching videos into a fine-grained sequence of knowledge intervals arranged in a preset learning order. This effectively compensates for the shortcomings of relying solely on page switching, which leads to excessively coarse or fragmented interval granularity. As a result, the final generated knowledge intervals possess both the cohesive integrity of teaching semantics and the reasonable duration of cognitive load.

[0078] In some specific implementations, multiple initial knowledge intervals are adjusted in conjunction with interval duration rules, including: In response to the initial knowledge interval satisfying the first condition, multiple initial knowledge intervals are merged into a candidate knowledge interval, wherein the first condition includes the duration of the initial knowledge interval being less than the minimum duration of the preset interval, and / or multiple initial knowledge intervals containing the same knowledge point; In response to the initial knowledge interval satisfying the second condition, the initial knowledge interval is split into multiple candidate knowledge intervals, wherein the second condition includes the duration of the initial knowledge interval being greater than the maximum preset interval duration, and / or the initial knowledge interval containing multiple independent knowledge points.

[0079] Here, the initial knowledge interval is a series of video segments initially cut from the learning video based on primary signals such as page switching timestamps. These are raw slices that have not been optimized for semantics and duration, and may be too fragmented or too long to be used directly as knowledge intervals.

[0080] Here, the common knowledge points refer to the core knowledge themes presented in different candidate intervals being consistent, without any substantial conceptual transformation.

[0081] Among these methods, identification can be achieved by the overlap of keywords in the text content within the interval, the semantic similarity of the audio text exceeding a threshold, or manually labeled with the same knowledge point.

[0082] Here, an independent knowledge point is a parallel knowledge unit within an initial knowledge interval that is semantically distinct and does not have a strong dependency relationship.

[0083] This can be achieved by detecting independent knowledge points through explicit switching phrases in the ASR text (such as "Next, let's look at another one") or obvious shifts in the text's theme.

[0084] Specifically, initial candidate segments are first formed using page-turning timestamps, and then merged or subdivided by combining PPT text titles, question numbers, knowledge point keywords, audio pauses, subtitle semantic density, and shortest / longest interval duration rules.

[0085] Specifically, when the interval between two adjacent page turns is less than 30 seconds and the PPT content is a transition, table of contents, or introductory page, the system automatically merges it into the previous interval to avoid the interval being too fragmented; when a single interval exceeds 15 minutes, or when there are multiple question numbers, example question start anchors, obvious audio pauses, or semantic shifts in subtitles on the same page, the system automatically subdivides the intervals.

[0086] For example, the system extracts candidate knowledge point titles (topics) and identifies the zone type (zone_type) by parsing the titles, chapter numbers, question numbers, keywords, and page layout in the PPT text, combined with the explanatory semantics in the audio-visual transcription subtitles. If the PPT structure is clear, the PPT chapter / question number is used as the zone boundary; if the PPT structure is unclear, the zone boundary is generated through subtitle topic clustering, audio pauses, and manual correction by the teacher.

[0087] In one embodiment, after splitting or merging, adjacent boundaries are fine-tuned to avoid abrupt interruptions to audio or video due to cutting. For example, the cutting point is aligned to the nearest silence gap or sentence end. Finally, all final intervals are checked to ensure they meet the duration constraints. If there are edge cases (such as the last interval failing to meet the minimum or maximum constraints due to special content), an exception exemption can be configured.

[0088] In one embodiment, the system dynamically adjusts T_max and T_min based on the historical average learning time and attention span data of the video's audience. For example, if it is observed that the exit rate rises sharply after most users watch for more than 10 minutes, T_max can be automatically reduced from 15 minutes to 10 minutes. Simultaneously, for sections that are frequently fast-forwarded or marked as "understood" by users, the system can automatically shorten the granularity of the knowledge interval for that section or mark it as skippable, achieving personalized structural optimization.

[0089] For example, suppose the video is titled "Complete Explanation of Linear Functions," with a preset minimum duration of 15 seconds and a maximum duration of 420 seconds, and 12 initial candidate knowledge intervals. The first condition for merging occurs when three candidate intervals exist: a 5-second slide flip-through interval, an 8-second concept blackboard writing interval, and a 6-second text annotation interval. All three explain the definition of a linear function, belonging to the same knowledge point, and their durations are all less than the minimum of 15 seconds, satisfying the first condition. The system merges these three intervals, generating a 19-second integrated candidate interval for the function definition. The second condition for splitting occurs when the merged interval generates a 450-second candidate interval for a major problem explanation, exceeding the maximum of 420 seconds. This interval contains three independent knowledge points: "question analysis," "formulation and solution," and "error correction," satisfying the second condition. The system, based on audio semantic breakpoints, splits this interval into two independent candidate intervals: a 220-second equation-solving interval and a 230-second error correction interval.

[0090] Thus, the interval adaptive adjustment scheme driven by the synergy of duration constraints and knowledge point semantics enables the learning video playback system to intelligently reorganize the initial knowledge intervals initially divided based on page switching. Through bidirectional dynamic control of merging and splitting, the final generated candidate knowledge intervals achieve dual optimization in both duration and semantic dimensions. This avoids frequent context switching caused by fragmentation and prevents learning fatigue caused by excessively long intervals, providing support for the semantic accuracy of subsequent breakpoint positioning, the fine granularity of mastery assessment, and the scientific nature of adaptive playback strategies.

[0091] In some specific implementations, the method further includes: In response to a user ending their viewing of a learning video, update and store the knowledge interval identifier and offset within the knowledge interval corresponding to the historical exit position; Collect user learning playback behavior data, which includes at least one of the following: effective viewing time of knowledge sections in learning videos, number of revisits, number of times you exited or entered the video, number of effective pauses, and number of times you watched the entire video. Based on the learning playback behavior data, update and store the learning level of the knowledge interval.

[0092] Here, the offset within the knowledge interval is the playback time displacement calculated from the starting position of the knowledge interval.

[0093] Here, effective viewing time is the cumulative time a user spends watching and can be considered as learning within a certain knowledge range.

[0094] Invalid viewing can include speeds below 0.5x, dragging to fast forward or skip, background suspension with mute, and paused content with no action.

[0095] Here, the number of times a user skips out is the number of times they actively skip the remaining content of a knowledge section by dragging the progress bar or clicking the next section; the number of times a user jumps in is the number of times they jump directly from another part of the video to the beginning of the current knowledge section.

[0096] Here, the learning level is a dynamic value recalculated and written back to storage based on the latest learning playback behavior data. It is a state variable that evolves over time, triggered by each viewing session, thus always maintaining the most up-to-date estimate of the user's current knowledge level.

[0097] Specifically, the system listens for the following events as trigger points for data collection and updates: when the user actively closes the player or exits the application, when the user switches to another video, or when the player enters the background for more than a certain period of time, it is determined to be a temporary departure. A snapshot is taken for temporary storage and a heartbeat is reported at regular intervals to prevent data loss due to accidental exit.

[0098] Specifically, the player continuously records user actions during the session and aggregates statistics based on knowledge intervals. The core fields collected include: interval entry / exit timestamps, playback speed change events, progress bar dragging events, and pause / play events.

[0099] Specifically, the platform pre-configures behavioral indicator weights, assigning positive weighting coefficients to positive indicators (number of complete views, number of revisits, effective viewing time, and number of effective pauses for reflection), and assigning deduction coefficients to negative indicators (number of bounces). The six behavioral data points, statistically analyzed in intervals, are substituted into the formula to output a standardized mastery score within the range of 0 to 100. For intervals with no historical records of first-time learning, the initial score is directly generated using the current behavioral data. For intervals with historical mastery scores, a smooth weighted update of the old and new scores is used to avoid significant fluctuations in scores caused by a single extreme operation.

[0100] For example, mastery level = interval completion rate (effective viewing time / total interval time) * 0.4 + active return visit rate (number of replays / total number of actions within the interval) * 0.25 - bounce rate (number of bounces / number of entries) * 0.15 + number of pauses for reflection (number of effective pauses) * 0.2 + number of complete playbacks (number of times the interval was watched in its entirety) * 0.3.

[0101] Thus, the dynamic mastery closed-loop update scheme based on breakpoint semantic recording and multi-dimensional behavioral evaluation enables the learning video playback system to simultaneously complete the structured persistence of breakpoint positions and the real-time refresh of knowledge interval mastery status when the user ends the current viewing. This breaks through the limitations of traditional schemes that only store a single physical timestamp and cannot reflect learning quality. The system can automatically distinguish between breakpoint storage and resume playback strategies based on learning quality, thereby ensuring high continuity and personalized adaptation capabilities across learning sessions.

[0102] In some specific implementations, in response to learning playback behavior data including the effective viewing time of knowledge intervals in the learning video, the collection of user learning playback behavior data includes: Obtain the user's learning duration factor for the knowledge interval; wherein, the learning duration factor includes at least one of continuous dwell time, replay time, jump-in dwell time, and jump-out time; The effective viewing time for a knowledge interval is obtained by weighting and fusing the learning duration factors using preset weights.

[0103] Here, continuous dwell time is the length of time a user can continuously watch within a knowledge section without any interruption such as jumping, dragging, or speeding up playback, and at a normal playback rate; replay time is the length of time a user can return to a previously played position within the current knowledge section by dragging the progress bar or clicking the rewind button and watch it again; jump-in dwell time is the length of time a user can jump directly from another position in the video to a specific position within the current section and continuously watch it at that position; and jump-out time is the length of time a user briefly stays within a knowledge section before performing a jump-out operation.

[0104] For example, effective viewing time = continuous viewing time + 1.5 × replay time + 0.5 × jump-in viewing time - 0.2 × jump-out time.

[0105] In this way, the undifferentiated time statistics are subdivided into multimodal behaviors such as continuous, replay, and jump, and differentiated weighted calculations are performed. This allows the effective viewing time to evolve from physical time to cognitive investment time, providing a precise quantitative input for mastery assessment that eliminates inflated figures and focuses on deep learning.

[0106] Thus, the effective viewing time calculation scheme based on the decoupling of learning behavior types and differentiated weighting performs weighted fusion based on the differences in cognitive input carried by various behaviors, transforming physical dwell time into effective viewing time that can truly reflect the quality of learning input. This effectively suppresses the interference of invalid time generated by low-quality operations on subsequent analysis, while amplifying positive signals and providing purer and more accurate input features for the mastery assessment model, thereby ensuring the accuracy of adaptive playback decisions and the credibility of learning quality diagnosis.

[0107] In some specific implementations, the method further includes: In response to the detection of related knowledge points, multiple knowledge areas are aggregated into chapters; Store the knowledge interval identifier, the offset within the knowledge interval, and the chapter identifier as the historical exit position.

[0108] Here, a chapter is a higher-level organizational unit that aggregates multiple knowledge areas according to the hierarchical logic of the knowledge system.

[0109] A chapter can include multiple levels, with multiple knowledge areas aggregated into a first-level chapter. Furthermore, multiple first-level chapters can be aggregated into a second-level chapter.

[0110] For example, Chapter 1 (i.e., Level 2 chapter) can contain several sections (i.e., Level 1 chapter). A section can be further divided into knowledge sections such as concept explanation section, example demonstration section, and summary and review section. Each knowledge section can cover one PPT page or multiple consecutive PPT pages. When a PPT page contains multiple examples, multiple sub-questions, or multiple obvious explanation stages, the system can split multiple knowledge sections on the same page based on question number, in-page animation / annotation time, subtitle semantic boundaries, or teacher supplementary anchor points.

[0111] Specifically, it supports the coexistence of chapter-level or section-level progress nodes and knowledge interval breakpoints at both granular levels. Chapter-level or section-level nodes are coarse-grained nodes obtained by aggregating multiple knowledge intervals, mainly used for recommending the next chapter, displaying overall progress, or planning cross-chapter learning paths; knowledge interval breakpoints are fine-grained replay units used to accurately locate a specific zone_id and its offset within the interval.

[0112] For example, "Chapter 1" can contain several sections, and each section contains multiple knowledge zones. A chapter corresponds to the upper-level aggregation range, but it does not mean that a chapter can only correspond to one zone. When locating breakpoints, the knowledge zone level zone_id is used first; when course-level recommendations require jumping to the next chapter or section, chapter-level or section-level aggregation nodes are used.

[0113] In one embodiment, chapter division can be performed before knowledge interval division, i.e., pre-defined chapter division.

[0114] In one embodiment, Figure 3 This is a schematic diagram of PPT-driven knowledge interval segmentation in an embodiment of this application, as shown below. Figure 3 As shown, the PPT-driven knowledge interval segmentation in this application includes: a hierarchical content architecture and an interval division rule system.

[0115] Specifically, the multi-level teaching content hierarchy includes a top-down progressive link from the top-level course to micro-breakpoints and an aggregation node branch link for progress control. The top-down progressive link from the top-level course to micro-breakpoints is the most basic content carrier, starting with the course / chapter. Taking Chapter 1 as an example, it is broken down to the section-level unit, such as 1.1 Conceptual Basics. The section-level unit is further divided into multiple independent knowledge intervals, labeled Z001, Z002, and Z003. Finally, each knowledge interval is mapped to a candidate breakpoint on a PPT page or within the page, serving as the time boundary for interval division, forming the top-down main line of course chapters, section-level units, knowledge intervals, and breakpoints within PPT pages. The aggregation node branch link system for progress control has a separate chapter-level aggregation node, whose core function is to support post-class content recommendations and global learning progress statistics. Chapter-level nodes aggregate downwards to generate section-level nodes. A single section-level node can accommodate multiple knowledge intervals of different types. The diagram divides these into three typical knowledge intervals: Z001 Concept Explanation Interval: Supports content spanning multiple PPT pages, bound to the corresponding PPT page-turning timestamps, where t1 is the starting page-turning time for the concept explanation and t2 is the page-turning time for the continuation of the concept content; Z002 Example Demonstration Interval: A single PPT page can hold independent breakpoints for multiple questions, corresponding to the start and switch time nodes of the two example questions q1 and q2 within page P3; Z003 Summary and Review Interval: Can merge transitional PPT pages, bound to the page-turning time t4 of page P4 as the starting boundary of the summary content.

[0116] Specifically, the complete knowledge interval division rule includes data sources covering multiple types of classroom behavior signals, such as time intervals generated by PPT page turning, breakpoints for question switching within a single PPT page, pauses in lecture audio, semantic segmentation points in subtitles, and demarcation points manually marked by teachers. Based on the above multi-source boundary data, a unified granularity control rule is implemented: segments shorter than 30 seconds are automatically merged, and when a single segment exceeds 15 minutes, it is further subdivided based on semantic density. At the same time, the same complete knowledge point is allowed to span multiple PPT pages, and a single PPT page is not mandatory as the upper limit for interval division. The granularity rule ultimately outputs standardized breakpoint matching rules, clearly defining a one-to-one correspondence between logical intervals and knowledge intervals. Each zone_id uniquely matches only one independent knowledge interval, while chapter and section levels only serve as upper-level aggregation and statistical nodes and do not participate in the division and binding of lower-level time intervals.

[0117] In one embodiment, Figure 4 This is a schematic diagram of the logic breakpoint detection and resume playback process in the embodiments of this application, such as... Figure 4As shown, the logical breakpoint detection and resume playback process in this application includes: When a user opens a course video, the system first parses the timestamp corresponding to the PPT page in the background, divides the knowledge intervals based on the timestamps, and simultaneously builds a completion model for evaluating learning progress. After preprocessing, the system determines whether the user has any historical breakpoint records: if there are no historical breakpoint records, the video plays directly from the beginning; if there are historical breakpoint records, the system retrieves the breakpoint status data, which includes the logical interval to which the breakpoint belongs and the playback offset within that interval. The system verifies the mastery level of the knowledge interval where the breakpoint is located and compares it with a preset threshold: if the mastery level is lower than the threshold, it means that the knowledge point is a weak point, and the video will directly jump to the logical starting point of this knowledge interval, while a pop-up window prompts the user about the weak knowledge point; if the mastery level is higher than or equal to the threshold, it will directly locate the last saved breakpoint time point and resume playback normally. Regardless of the jump playback method used, the system will collect various viewing behavior data of the user in real time throughout the video playback process, and continuously update the heat weight and learning completion value of each knowledge interval based on the behavior data. After each playback session, the system re-verifies whether the user's mastery of the current knowledge segment meets the required standard. If not, the segment is marked as a weak segment and added to the knowledge point review recommendation queue. If the standard is met, the segment is marked as mastered, and the user's overall learning progress map is updated accordingly. The system uniformly saves the logical breakpoints formed during this playback session, storing the segment's unique identifier and playback offset within the segment. This provides historical data support for the user's next video viewing, forming a complete learning progress cycle management process.

[0118] In another embodiment, the logical breakpoint detection and resume playback process in this application further includes: after the video is uploaded, the system backend automatically extracts the PPT page-turning timeline, cuts knowledge intervals, and constructs an interval metadata table; when the user plays the video, the behavior collection SDK reports events such as pause / jump in / jump out / pause / replay in real time; behavior events trigger heatmap weight calculation, and update the completion CR value of each knowledge interval in real time; when the user pauses or exits, the system automatically saves the logical breakpoint status (interval ID + interval offset + CR + mastery score); when the user re-enters the course, the system queries the breakpoint status, executes the resume playback starting point positioning algorithm, and decides whether to resume playback from the logical starting point or the offset; during playback, the mastery score is continuously updated, and after completing an interval, it is determined whether to add it to the weak knowledge backtracking queue; when the user views the "learning progress map", the system renders a heatmap according to the interval completion, and displays the mastery status of the entire course in three colors.

[0119] Example 1 When a student watches for 20 minutes within the "Chapter 4: Classes and Objects" knowledge section, skipping the first 10 minutes and entering playback from the middle of the section, then exiting the video after only watching the latter 10 minutes, the system records the effective viewing time within that section as 600 seconds. The first half of the content skipped by the user is not counted as effective viewing time. The calculated completion rate (CR) for this section is 600 / 1200 = 0.50, which is below the threshold of 0.75. The system then further calculates the completion rate based on a multi-dimensional weighting formula that considers effective viewing, replay, exiting, and pausing. The calculated mastery score is 0.50*40+0*25+0*(-15)+2*20+0*30=60 points, which is still below the passing threshold. When the student opens the video again, the system will directly jump to the logical starting point of the "Classes and Objects" knowledge section and pop up a prompt informing the student that the knowledge point was not viewed completely and suggesting that they learn it from the beginning. After the student completely relearns the entire knowledge section, the section completion rate increases to 0.91, the mastery score is updated to 82 points, and the knowledge point is simultaneously changed to the mastered status.

[0120] Example 2 After completing all courses, students scored 45 and 31 points respectively on the "Chapter 3: Function Definition" and "Chapter 6: Module Import" knowledge areas. The system automatically added these two areas to the weak knowledge review queue and pushed a notification on the course homepage, informing students that there were two knowledge points that needed to be reinforced. After students clicked the notification, the system generated a recommendation list based on their mastery scores from low to high, prioritizing Chapter 6 (31 points) and then Chapter 3 (45 points). When students selected to study Chapter 6, the player directly jumped to the beginning of that knowledge area. After the students finished studying, the mastery score for that area was updated to 78 points, and the system simultaneously refreshed the learning progress heatmap, changing the marker for Chapter 6 from red (representing weakness) to orange. The weak knowledge point recommendation list was also updated accordingly.

[0121] Example 3 Students learn on their mobile phones during their ten-minute commute in the morning. The playback pauses at the eighth minute of the "Chapter 2: Control Flow" knowledge section, at which point the completion rate (CR) is 0.60. The system saves the breakpoint information, recording the section number as Z002 and the offset duration as 480 seconds. In the evening, students switch to their computers to continue learning. The system retrieves historical breakpoint data and determines that the completion rate of 0.60 is below the passing threshold of 0.75. It then triggers the logical start point continuation strategy, skipping the video from the previous pause position at 480 seconds and jumping directly to the beginning of the "Control Flow" section. A prompt appears informing the student that they failed to fully learn this knowledge point last time and suggesting a review from the beginning. After combining the ten-minute commute learning time with the thirty-minute evening learning time, the completion rate of this chapter section increases to 0.95, marking it as mastered, and the corresponding block on the learning progress map is updated to green. The core advantage of this solution is that cross-device breakpoints rely on semantic knowledge section numbers for storage, eliminating limitations on the playback terminal and ensuring seamless synchronization of progress across mobile phones, computers, and tablets.

[0122] It should be understood that, although Figures 1-4 The steps in the flowchart are shown sequentially as indicated by the arrows, but these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order in which these steps are executed, and they can be performed in other orders. Figures 1-4 At least some of the steps in the process may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be executed in turn or alternately with other steps or at least some of the sub-steps or stages of other steps.

[0123] In one embodiment, such as Figure 5As shown, a learning video playback device is provided. The device includes: a first processing module 501, used to acquire learning videos and divide the learning videos into multiple knowledge intervals arranged according to a preset learning order based on teaching units; a second processing module 502, used to receive a user's playback command for the learning videos, parse and acquire the historical exit playback position corresponding to the playback command; a third processing module 503, used to determine candidate knowledge intervals from multiple knowledge intervals based on the historical exit playback position, wherein the candidate knowledge intervals include a preceding knowledge interval and a current knowledge interval, the current knowledge interval represents the knowledge interval corresponding to the historical exit playback position, and the preceding knowledge interval represents the knowledge interval arranged before the current knowledge interval; a fourth processing module 504, used to collect and determine the heat features of the candidate knowledge intervals based on preset heat feature rules; a fifth processing module 505, used to determine the user's learning level of the candidate knowledge intervals by combining the heat features of the candidate knowledge intervals; and a sixth processing module 506, used to determine the preceding knowledge interval or the current knowledge interval as the target knowledge interval based on the learning level of the candidate knowledge intervals, and play the target knowledge interval.

[0124] As a preferred implementation, in this embodiment of the application, the sixth processing module 506 is specifically used to: in response to detecting that the mastery of a prior knowledge interval is less than or equal to the mastery threshold, take the first prior knowledge interval whose mastery is less than or equal to the mastery threshold according to the preset learning order as the target knowledge interval; in response to the mastery of a prior knowledge interval being greater than the mastery threshold, take the current knowledge interval as the target knowledge interval.

[0125] As a preferred implementation, in this embodiment of the application, the sixth processing module 506 is further configured to: in response to the preceding knowledge interval being the target knowledge interval, start playback from the starting position of the preceding knowledge interval; or, in response to the current knowledge interval being the target knowledge interval, determine the interval type of the current knowledge interval, determine the playback position of the current knowledge interval based on the interval type, and start playback from the playback position.

[0126] As a preferred implementation, in this embodiment of the application, the sixth processing module 506 is further configured to: respond to the fact that the interval type of the current knowledge interval is theoretical explanation, start playing from the beginning position of the current knowledge interval; respond to the fact that the interval type of the current knowledge interval is example explanation, determine the example segment corresponding to the historical exit position, and start playing from the beginning position of the example segment.

[0127] As a preferred implementation, in this embodiment of the application, the first processing module 501 is specifically used to: parse the learning video to obtain the text content, audio explanation, and page switching timestamps in each display interface of the learning video; divide the learning video into multiple initial knowledge intervals according to the page switching timestamps; adjust the multiple initial knowledge intervals according to the audio explanation and text content, combined with the interval duration rules, to determine multiple knowledge intervals of the learning video, and arrange the multiple knowledge intervals according to a preset learning order.

[0128] In a preferred embodiment of this application, the first processing module 501 is further configured to: merge multiple initial knowledge intervals into a candidate knowledge interval in response to the initial knowledge interval satisfying a first condition, wherein the first condition includes the duration of the initial knowledge interval being less than the minimum duration of a preset interval, and / or multiple initial knowledge intervals containing the same knowledge point; and split the initial knowledge interval into multiple candidate knowledge intervals in response to the initial knowledge interval satisfying a second condition, wherein the second condition includes the duration of the initial knowledge interval being greater than the maximum duration of a preset interval, and / or the initial knowledge interval containing multiple independent knowledge points.

[0129] In a preferred embodiment of this application, the device further includes a seventh processing module, which is specifically used for: responding to the user ending the viewing of the learning video, updating and storing the knowledge interval identifier and the offset within the knowledge interval corresponding to the historical exit position; collecting the user's learning playback behavior data, wherein the learning playback behavior data includes at least one of the effective viewing duration, number of revisits, number of exits, number of entries, number of effective pauses, and number of complete views of the knowledge interval in the learning video; and updating and storing the learning level of the knowledge interval based on the learning playback behavior data.

[0130] As a preferred implementation, in this embodiment of the application, the seventh processing module is further configured to: obtain the learning duration factor of the user for the knowledge interval; wherein, the learning duration factor includes at least one of continuous dwell time, replay time, jump-in dwell time and jump-out time; and use preset weights to perform weighted fusion of the learning duration factor to obtain the effective viewing time of the knowledge interval.

[0131] For specific limitations on the learning video playback device, please refer to the limitations on the learning video playback method mentioned above, which will not be repeated here.

[0132] In one embodiment, an electronic device is provided, which may be a target bus interface device, and its internal structure diagram may be as follows: Figure 6As shown, the electronic device includes a processor, memory, network interface, display screen, and input devices connected via a system bus. The processor provides computing and control capabilities. The memory includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage medium. The network interface is used to communicate with external terminals via a network connection. When the computer program is executed by the processor, it implements a learning video playback method. The display screen can be a liquid crystal display (LCD) or an e-ink display. The input devices can be a touch layer covering the display screen, buttons, a trackball, or a touchpad mounted on the device's casing, or an external keyboard, touchpad, or mouse.

[0133] Embodiments of this application also provide a computer-readable storage medium storing a computer program, wherein the computer program is configured to execute the steps in any of the above embodiments of the learning video playback method when it is run.

[0134] In one exemplary embodiment, the aforementioned computer-readable storage medium may include, but is not limited to, various media capable of storing computer programs, such as a USB flash drive, read-only memory (ROM), random access memory (RAM), portable hard disk, magnetic disk, or optical disk.

[0135] Embodiments of this application also provide a computer program product, which includes a computer program that, when executed by a processor, implements the steps in any of the above embodiments of the learning video playback method.

[0136] Embodiments of this application also provide another computer program product, including a non-volatile computer-readable storage medium storing a computer program, which, when executed by a processor, implements the steps in any of the above embodiments of the learning video playback method.

[0137] Any of the components, modules, units, parts, methods, and operations described herein can be implemented using software, firmware, hardware (e.g., fixed logic circuitry), manual processing, or any combination thereof. Alternatively or additionally, any functionality described herein can be executed at least in part by one or more hardware logic components, such as, but not limited to, a central processing unit (CPU), a field-programmable gate array (FPGA), an application-specific integrated circuit (ASIC), an application-specific standard product (ASSP), a system-on-a-chip (SoC), a complex programmable logic device (CPLD), a microprocessor (MCU), etc. The terms "system," "computing device," or "apparatus" as used herein encompass various means, devices, and machines for processing data, including, for example, one or more programmable processors, computers, SoCs, or combinations thereof. The apparatus may also include code that creates an execution environment for the computer program in question, such as code constituting processor firmware, a protocol stack, a database management system, an operating system, a cross-platform runtime environment, a virtual machine, or one or more combinations thereof. The aforementioned computer program (also known as a program, software, software application, app, script, or code) can be written in any form of programming language, including compiled or interpreted languages, declarative or procedural languages, and can be deployed in any form, including as a standalone program or as a module, component, subroutine, object, or other unit suitable for a computing environment.

[0138] Those skilled in the art will further recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0139] The foregoing has provided a detailed description of a learning video playback method, apparatus, and electronic device provided in this application. Specific examples have been used to illustrate the principles and implementation methods of this application. The descriptions of the embodiments above are only intended to aid in understanding the method and core ideas of this application. It should be noted that those skilled in the art can make various improvements and modifications to this application without departing from its principles, and these improvements and modifications also fall within the protection scope of the claims of this application.

Claims

1. A method for playing learning videos, characterized in that, The method includes: Acquire learning videos and divide the learning videos into multiple knowledge intervals arranged according to a preset learning order based on teaching units; Receive the user's playback command for the learning video, parse and obtain the historical exit position corresponding to the playback command; Candidate knowledge intervals are determined from the plurality of knowledge intervals based on the historical exit playback position, wherein the candidate knowledge intervals include the preceding knowledge intervals and the current knowledge interval, the current knowledge interval represents the knowledge interval corresponding to the historical exit playback position, and the preceding knowledge interval represents the knowledge interval arranged before the current knowledge interval; Based on preset thermal feature rules, thermal features of the candidate knowledge intervals are collected and determined. Based on the heat map characteristics of the candidate knowledge intervals, the user's learning level of the candidate knowledge intervals is determined; Based on the learning level of the candidate knowledge intervals, the preceding knowledge interval or the current knowledge interval is determined as the target knowledge interval, and the target knowledge interval is played.

2. The learning video playback method according to claim 1, characterized in that, The learning level includes mastery; determining the prior knowledge interval or the current knowledge interval as the target knowledge interval based on the learning level of the candidate knowledge interval includes: In response to the detection that there exists a prior knowledge interval where the mastery level is less than or equal to the mastery level threshold, the first prior knowledge interval where the mastery level is less than or equal to the mastery level threshold, according to the preset learning order, is taken as the target knowledge interval. In response to the fact that the mastery of the prior knowledge interval is greater than the mastery threshold, the current knowledge interval is taken as the target knowledge interval.

3. The learning video playback method according to claim 2, characterized in that, The playback of the target knowledge range includes: In response to the premise knowledge interval being the target knowledge interval, playback begins from the starting position of the premise knowledge interval; or, In response to the current knowledge interval being the target knowledge interval, the interval type of the current knowledge interval is determined, and the playback position of the current knowledge interval is determined based on the interval type, and playback begins from the playback position.

4. The learning video playback method according to claim 3, characterized in that, The interval type includes theoretical explanations and example explanations; determining the playback position of the current knowledge interval based on the interval type includes: In response to the fact that the current knowledge interval is of the interval type of theoretical explanation, playback will begin from the beginning of the current knowledge interval; In response to the interval type of the current knowledge interval, the example problem explanation is used to determine the example problem segment corresponding to the historical exit playback position, and playback starts from the beginning position of the example problem segment.

5. The learning video playback method according to claim 1, characterized in that, The process of acquiring learning videos and dividing them into multiple knowledge segments arranged according to a preset learning order by teaching units includes: The learning video is analyzed to obtain the text content, audio explanations, and page switching timestamps of each display interface in the learning video. Based on the page switching timestamp, the learning video is divided into multiple initial knowledge intervals; Based on the audio explanation and the text content, the multiple initial knowledge intervals are adjusted according to the interval duration rules to determine multiple knowledge intervals of the learning video. The multiple knowledge intervals are arranged according to a preset learning order.

6. The learning video playback method according to claim 5, characterized in that, The adjustment of the multiple initial knowledge intervals based on the combined interval duration rule includes: In response to the initial knowledge interval satisfying the first condition, the plurality of initial knowledge intervals are merged into a candidate knowledge interval, wherein the first condition includes the duration of the initial knowledge interval being less than the minimum duration of a preset interval, and / or the plurality of initial knowledge intervals containing the same knowledge point; In response to the initial knowledge interval satisfying the second condition, the initial knowledge interval is split into multiple candidate knowledge intervals, wherein the second condition includes the duration of the initial knowledge interval being greater than the maximum duration of a preset interval, and / or the initial knowledge interval containing multiple independent knowledge points.

7. The learning video playback method according to any one of claims 1 to 6, characterized in that, The method further includes: In response to the user ending the viewing of the learning video, update and store the knowledge interval identifier and the offset within the knowledge interval corresponding to the historical exit position; Collect the user's learning playback behavior data, wherein the learning playback behavior data includes at least one of the following: effective viewing time of knowledge sections in the learning video, number of revisits, number of exits, number of entries, number of effective pauses, and number of complete views; Based on the learning playback behavior data, update and store the learning level of the knowledge interval.

8. The learning video playback method according to claim 7, characterized in that, In response to the learning playback behavior data including the effective viewing time of knowledge sections in the learning video, the collection of the user's learning playback behavior data includes: Obtain the learning duration factor of the user for the knowledge interval; wherein, the learning duration factor includes at least one of continuous dwell time, replay time, jump-in dwell time, and jump-out time; The effective viewing time of the knowledge interval is obtained by weighting and fusing the learning duration factors using preset weights.

9. A learning video playback device, characterized in that, The device includes: The interactive module is used to receive user commands to play the learning videos; A computing module includes a processor and a memory; the computing module is connected to the interaction module and is used to implement the learning video playback method as described in any one of claims 1-8.

10. An electronic device, characterized in that, The electronic device includes: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor, when executing the computer program, implements the steps of the method according to any one of claims 1 to 8; A display device is connected to the learning video playback device and is used to display and play learning videos.

Images