VR commodity focus recognition method based on eyeball tracking

Abstract

The invention discloses a VR commodity focus recognition method and system based on eyeball tracking. The method comprises the steps that VR equipment integrated with a Tobii eye tracker is used for collecting coordinates of a fixation point of a user in real time; inputting the collected coordinates of the fixation point into a convolutional neural network model, recognizing a fixation target through the convolutional neural network model, and judging whether the fixation target is a commodity or not; and if the gazing target is the commodity, counting the total duration of gazing the commodity by the user, obtaining a commodity attention ranking result based on the total duration, and adjusting the virtual shelf layout according to the ranking result. According to the invention, a complete user attention perception and business conversion solution is provided for VR e-commerce, and virtual shopping is promoted to develop towards the intelligent and personalized direction.

Description

Technical Field

[0001] The present invention relates to the field of human-computer interaction technology, and in particular to a VR product focus recognition method based on eye tracking. Background Art

[0002] With the rapid development of science and technology, virtual reality (VR) technology is gradually gaining widespread application in various fields. VR shopping, as an emerging shopping model, is gaining increasing attention from consumers. Traditional e-commerce shopping models primarily display products through a two-dimensional interface, with relatively limited user interaction methods, making it difficult to provide an immersive shopping experience. VR technology, however, can create a highly realistic virtual shopping environment, allowing users to freely browse and select products as if they were in a real mall or store, greatly enhancing the immersive and enjoyable shopping experience.

[0003] While there has been some research on VR shopping and eye-tracking technology, many challenges remain, including accurate and real-time product focus recognition and how to better leverage user attention data to optimize the shopping experience. For example, how to accurately distinguish products from other elements (such as user interfaces) in complex VR scenes, how to deal with gaze data noise caused by factors such as eye tremors, and how to dynamically adjust virtual shelf layouts based on user attention to improve product exposure and sales conversion rates—all of these issues require further research and resolution. Therefore, developing a VR product focus recognition method based on eye tracking has important practical significance and application value. Summary of the Invention

[0004] The purpose of this section is to summarize some aspects of the embodiments of the present invention and briefly introduce some preferred embodiments. Some simplifications or omissions may be made in this section and the abstract and title of this application to avoid obscuring the purpose of this section, the abstract and the title of the invention, and such simplifications or omissions should not be used to limit the scope of the present invention.

[0005] In light of the above-mentioned existing problems, the present invention proposes a VR product focus recognition method based on eye tracking to address the difficulty in accurately identifying the user's product focus in VR shopping scenarios and dynamically optimize virtual shelf layout to improve the shopping experience and sales efficiency.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0007] In a first aspect, the present invention provides a method for identifying VR product focus based on eye tracking, comprising:

[0008] Use VR devices integrated with Tobii eye trackers to collect user gaze coordinates in real time;

[0009] Inputting the collected gaze point coordinates into a convolutional neural network model, identifying the gaze target through the convolutional neural network model, and determining whether the gaze target is a product;

[0010] If the gaze target is a product, the total time the user gazes at the product is counted, a product attention ranking result is obtained based on the total time, and the virtual shelf layout is adjusted according to the ranking result.

[0011] As a preferred solution of the VR product focus recognition method based on eye tracking described in the present invention, the user's gaze point coordinates are collected in real time using a VR device integrated with a Tobii eye tracker, including:

[0012] Continuously collect the pupil center coordinates and corneal reflection point coordinates of the user's left and right eyes at a sampling frequency of 50Hz;

[0013] Based on the pupil center coordinates and the corneal reflection point coordinates, the user's line of sight direction vector is calculated using the Purkinje imaging principle;

[0014] Mapping the sight line direction vector to the three-dimensional space coordinate system of the VR scene, and determining the coordinates of the intersection of the sight line direction vector and the virtual product plane as the coordinates of the user's gaze point;

[0015] The continuously collected gaze point coordinates are processed by Kalman filtering to eliminate the noise interference caused by eye tremors.

[0016] As a preferred solution of the VR product focus recognition method based on eye tracking described in the present invention, the convolutional neural network model includes a feature extraction layer, a feature classification layer, and an attention mechanism layer;

[0017] The feature extraction layer extracts image features of a preset pixel area around the gaze point, and the preset pixel area is dynamically adjusted according to the field of view and resolution of the VR device;

[0018] Mapping the image features to a predefined target category space using a feature classification layer, where the target categories include product categories and UI interface categories;

[0019] The image features extracted by the attention mechanism layer are then weighted to enhance the saliency of product features and suppress the interference of UI interface features.

[0020] The convolutional neural network model is pre-trained using a transfer learning method, with the ImageNet dataset as the base training set, and then fine-tuned using product images and UI interface images in VR shopping scenarios.

[0021] As a preferred embodiment of the VR product focus recognition method based on eye tracking described in the present invention, the method includes: identifying the gaze target by the convolutional neural network model and determining whether the gaze target is a product, including:

[0022] When the coordinates of the user's gaze point change, a local image block centered at the gaze point in the current frame of the VR scene image is captured;

[0023] Inputting the local image block into a convolutional neural network model to obtain a classification probability distribution output by the model;

[0024] If the classification probability of the commodity category exceeds the classification preset threshold, the gaze target is determined to be a commodity;

[0025] If the classification results in consecutive frames are all UI interface categories, the current gaze point acquisition and recognition process is suspended until a change in the gaze point coordinates is detected.

[0026] As a preferred solution of the VR product focus identification method based on eye tracking of the present invention, obtaining the product attention ranking result based on the total duration includes:

[0027] Create an attention record table for each product, recording the start time, end time, and duration of the product being looked at. Combine multiple attention events for the same product. If the time interval between two attention events is less than a preset interval threshold, the two attention durations are accumulated.

[0028] Calculate the total gaze time of each product and sort them from high to low according to the total gaze time to obtain the product attention ranking results;

[0029] The total gaze time is normalized and converted into an attention score of 0-100 to quantify the degree of attention received by the displayed products.

[0030] As a preferred solution of the VR product focus recognition method based on eye tracking described in the present invention, wherein: adjusting the virtual shelf layout according to the sorting result includes:

[0031] Move the top 20% of products to the first visual area of ​​the virtual shelf;

[0032] Dynamically adjust product display size based on product attention score;

[0033] A virtual shelf layout optimization model is established, with product click-through rate and conversion rate as the objective function, and the product display location and display method are iteratively optimized through reinforcement learning algorithm.

[0034] As a preferred solution of the VR product focus recognition method based on eye tracking described in the present invention, the total duration includes:

[0035] If the total duration of a user's continuous gaze on an ad product exceeds the preset time threshold, the advertising billing mechanism is triggered.

[0036] In a second aspect, the present invention provides a system for VR product focus recognition based on eye tracking, comprising:

[0037] A data module is used to collect the user's gaze coordinates in real time using a VR device integrated with a Tobii eye tracker;

[0038] A judgment module, configured to input the collected gaze point coordinates into a convolutional neural network model, identify the gaze target through the convolutional neural network model, and determine whether the gaze target is a commodity;

[0039] The adjustment module is used to count the total time the user gazes at the product if the gaze target is a product, obtain a product attention ranking result based on the total time, and adjust the virtual shelf layout according to the ranking result.

[0040] In a third aspect, the present invention provides a computing device, comprising:

[0041] memory and processor;

[0042] The memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions. When the computer-executable instructions are executed by the processor, the steps of the VR product focus recognition method based on eye tracking are implemented.

[0043] In a fourth aspect, the present invention provides a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, implement the steps of the eye-tracking-based VR product focus recognition method.

[0044] Compared with the existing technology, the present invention has the following beneficial effects: through high-frequency sampling and filtering processing, it effectively eliminates eye tremor noise, ensuring the real-time and accuracy of focus recognition; the innovative design of the attention mechanism layer combined with transfer learning effectively distinguishes between products and UI interfaces, reducing the misjudgment rate; the attention ranking based on total gaze time and dynamic optimization of shelves improve the click-through rate and conversion rate of products; the intelligent advertising billing mechanism is triggered based on the user's real attention, improving the effectiveness of advertising delivery; the local image block processing and reinforcement learning optimization of the adaptive field of view angle support the real-time layout adjustment of large-scale product libraries. This invention provides a complete user attention perception and business conversion solution for VR e-commerce, promoting the development of virtual shopping in the direction of intelligence and personalization. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. Those skilled in the art can also derive other drawings based on these drawings without inventive effort. Among them:

[0046] Figure 1 This is a schematic diagram of a general flow chart of a method for VR product focus recognition based on eye tracking according to an embodiment of the present invention. DETAILED DESCRIPTION

[0047] To make the above-mentioned objects, features, and advantages of the present invention more clearly understood, the following detailed description of the specific embodiments of the present invention is given in conjunction with the accompanying drawings. It is obvious that the described embodiments are only part of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by ordinary persons in this field without creative work should fall within the scope of protection of the present invention.

[0048] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0049] Secondly, the term "one embodiment" or "embodiment" herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in various places throughout this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive of other embodiments.

[0050] The present invention is described in detail with reference to schematic diagrams. For ease of illustration, cross-sectional views of device structures may be partially enlarged and not to scale when describing embodiments of the present invention. Furthermore, the schematic diagrams are merely illustrative and should not limit the scope of the present invention. Furthermore, in actual production, the three-dimensional dimensions of length, width, and depth should be included.

[0051] Furthermore, in the description of the present invention, it should be noted that the terms "upper, lower, inner, and outer" and other references to orientations or positional relationships are based on the orientations or positional relationships shown in the accompanying drawings and are intended solely to facilitate and simplify the description of the present invention. They are not intended to indicate or imply that the systems or components referred to must have, be constructed, or operate in a specific orientation, and therefore should not be construed as limitations on the present invention. Furthermore, the terms "first, second, or third" are used for descriptive purposes only and should not be construed as indicating or implying relative importance.

[0052] In this disclosure, unless otherwise specified or limited, the terms "mounted," "connected," and "connected" should be interpreted broadly. For example, they may refer to fixed, removable, or integral connections. They may also refer to mechanical, electrical, or direct connections, indirect connections through an intermediary, or internal communication between two components. Those skilled in the art will understand the specific meanings of these terms in this disclosure.

[0053] Example 1

[0054] Reference Figure 1 , as one embodiment of the present invention, provides a VR product focus recognition method based on eye tracking, comprising:

[0055] S100: Uses a VR device with an integrated Tobii eye tracker to collect the user's gaze coordinates in real time;

[0056] S102: Inputting the collected gaze point coordinates into a convolutional neural network model, identifying the gaze target through the convolutional neural network model, and determining whether the gaze target is a product;

[0057] S104: If the gaze target is a product, the total time the user gazes at the product is counted, a ranking result of product attention is obtained based on the total time, and the virtual shelf layout is adjusted according to the ranking result.

[0058] Furthermore, a VR device integrated with a Tobii eye tracker is used to collect the user's gaze coordinates in real time, including:

[0059] Continuously collect the pupil center coordinates and corneal reflection point coordinates of the user's left and right eyes at a sampling frequency of 50Hz;

[0060] Based on the pupil center coordinates and corneal reflection point coordinates, the user's line of sight direction vector is calculated using the Purkinje imaging principle;

[0061] Map the gaze direction vector to the three-dimensional space coordinate system of the VR scene, and determine the coordinates of the intersection of the gaze direction vector and the virtual product plane as the coordinates of the user's gaze point;

[0062] The continuously collected gaze point coordinates are processed by Kalman filtering to eliminate the noise interference caused by eye tremors.

[0063] It should be noted that by continuously collecting the coordinates of the pupil center and corneal reflection point at high frequency, combining the Purkinje imaging principle to calculate the gaze direction vector, and mapping it to VR three-dimensional space to determine the gaze point coordinates, and then using Kalman filtering to eliminate eye tremor noise, high-precision real-time tracking of user gaze behavior is achieved. This technical solution effectively improves the stability and accuracy of gaze point positioning, ensuring accurate capture of the user's gaze focus in dynamic VR scenes, providing a reliable data foundation for subsequent product identification and interaction, significantly reducing misjudgments and interference caused by physiological tremors or scene changes, and enhancing the system's ability to perceive the user's true attention intentions.

[0064] Furthermore, the convolutional neural network model includes a feature extraction layer, a feature classification layer, and an attention mechanism layer;

[0065] The feature extraction layer extracts image features of a preset pixel area around the gaze point. The preset pixel area is dynamically adjusted according to the field of view and resolution of the VR device.

[0066] Use the feature classification layer to map image features to predefined target category spaces, including product categories and UI interface categories;

[0067] The image features extracted by the attention mechanism layer are then weighted to enhance the saliency of product features and suppress the interference of UI interface features.

[0068] The convolutional neural network model is pre-trained using the transfer learning method, with the ImageNet dataset as the basic training set, and then fine-tuned using product images and UI interface images in VR shopping scenarios.

[0069] It should be noted that by dynamically adjusting the preset pixel area to extract image features, combining multi-category mapping with attention mechanism weighted processing, and using transfer learning to optimize model training, accurate classification and recognition of products and UI interfaces in VR scenes are achieved. This solution adapts to different VR device parameters, effectively focusing on the key features of the user's gaze area, enhancing the expressiveness of product features, while suppressing the interference of interface elements, and significantly improving the model's ability to understand complex shopping scenarios. The training strategy that combines pre-training and fine-tuning enables the model to converge quickly with limited samples, retaining the recognition ability of general image features while accurately adapting to the specific scene requirements of VR shopping. Ultimately, it achieves efficient and accurate classification of user gaze targets, providing a reliable basis for subsequent product attention analysis.

[0070] Furthermore, the gaze target is identified through a convolutional neural network model to determine whether the gaze target is a commodity, including:

[0071] When the coordinates of the user's gaze point change, a local image block centered at the gaze point in the current frame of the VR scene image is captured;

[0072] Input the local image block into the convolutional neural network model to obtain the classification probability distribution of the model output;

[0073] If the classification probability of the commodity category exceeds the preset classification threshold, the gaze target is determined to be a commodity;

[0074] If the classification results in consecutive frames are all UI interface categories, the current gaze point acquisition and recognition process is suspended until a change in the gaze point coordinates is detected.

[0075] Furthermore, the ranking results of product attention are obtained based on the total duration, including:

[0076] Create an attention record table for each product, recording the start time, end time, and duration of the product being looked at. Combine multiple attention events for the same product. If the time interval between two attention events is less than a preset interval threshold, the two attention durations are accumulated.

[0077] Calculate the total gaze time of each product and sort them from high to low according to the total gaze time to obtain the product attention ranking results;

[0078] The total gaze time is normalized and converted into an attention score of 0-100 to quantify the degree of attention received by the displayed products.

[0079] It should be noted that through gaze-driven local image block processing and a dynamic threshold judgment mechanism, combined with a multi-frame continuous recognition strategy, real-time and accurate classification of the user's gaze target is achieved, effectively reducing invalid calculations and misjudgment interference. At the same time, based on time series attention records and intelligent merging algorithms, it accurately captures users' sustained attention behavior and quantifies the degree of attention received by products through a normalized scoring system, forming a complete framework for quantitative analysis of user attention. This solution not only ensures the timeliness and accuracy of product identification, but also reduces system resource consumption through an intelligent sleep mechanism. Ultimately, it provides a reliable basis for the dynamic optimization of virtual shelves and enables in-depth exploration and visualization of users' potential shopping interests.

[0080] Furthermore, the virtual shelf layout is adjusted according to the sorting results, including:

[0081] Move the top 20% of products to the first visual area of ​​the virtual shelf;

[0082] Dynamically adjust product display size based on product attention score;

[0083] A virtual shelf layout optimization model is established, with product click-through rate and conversion rate as the objective function, and the product display location and display method are iteratively optimized through reinforcement learning algorithm.

[0084] Furthermore, the total duration includes:

[0085] If the total duration of a user's continuous gaze on an ad product exceeds the preset threshold, the advertising billing mechanism is triggered;

[0086] Specifically, the advertising billing mechanism is based on an attention value assessment model. This model, through multi-dimensional data fusion and intelligent decision-making algorithms, establishes a precise advertising effectiveness quantification system. The system first extracts features from user gaze behavior, categorizing it into three levels: browsing (short-duration gaze), interest (medium-duration gaze), and intent (long-duration gaze). These are then assigned different billing weights, enabling refined pricing of attention resources. A contextual analysis module is also introduced, combining data such as user shopping history, product interactions, and gaze trajectory consistency to dynamically assess the commercial conversion potential of ad gaze behavior and effectively filter out ineffective impressions. To adapt to the needs of diverse marketing scenarios, the system employs a dynamic threshold adjustment mechanism, adjusting the gaze duration threshold for triggering billing in real time based on factors such as the ad spot's prime visual location, product category attributes, and promotional activity status. To ensure a balanced balance of interests among all parties, the system incorporates a built-in anti-spam algorithm that assesses the credibility of user dwell behavior through gaze quality scores to prevent malicious clicks. The mechanism also provides a real-time data synchronization interface for seamless integration with third-party advertising settlement systems, enabling precise billing based on effective attention duration and generating multi-dimensional performance analysis reports. By converting users' real attention into quantifiable advertising value units, this solution builds an attention economy ecosystem in VR shopping scenarios, which not only improves the efficiency of advertising delivery but also ensures the smoothness and fairness of the user experience.

[0087] It should be noted that high-frequency sampling and filtering effectively eliminate eye tremor noise, ensuring real-time and accurate focus recognition. The innovatively designed attention mechanism layer, combined with transfer learning, effectively distinguishes between products and UI interfaces, reducing misjudgment rates. Attention ranking and shelf dynamic optimization based on total gaze duration improve product click-through rates and conversion rates. An intelligent advertising billing mechanism is triggered based on the user's true attention level, enhancing advertising effectiveness. Local image block processing with adaptive field of view and reinforcement learning optimization supports real-time layout adjustments for large-scale product libraries. This invention provides a complete user attention perception and commercial conversion solution for VR e-commerce, driving the development of virtual shopping towards intelligent and personalized directions.

[0088] The above is a schematic diagram of the eye-tracking-based VR product focus recognition method of this embodiment. It should be noted that the technical solution of the eye-tracking-based VR product focus recognition system and the technical solution of the eye-tracking-based VR product focus recognition method described above are based on the same concept. For details not described in detail in the technical solution of the eye-tracking-based VR product focus recognition system in this embodiment, please refer to the description of the technical solution of the eye-tracking-based VR product focus recognition method described above.

[0089] The VR product focus recognition system based on eye tracking in this embodiment includes:

[0090] A data module is used to collect the user's gaze coordinates in real time using a VR device integrated with a Tobii eye tracker;

[0091] A judgment module is used to input the collected gaze point coordinates into a convolutional neural network model, identify the gaze target through the convolutional neural network model, and determine whether the gaze target is a product;

[0092] The adjustment module is used to count the total time the user gazes at the product if the gaze target is a product, obtain the product attention ranking result based on the total time, and adjust the virtual shelf layout according to the ranking result.

[0093] This embodiment further provides a computing device suitable for VR product focus recognition based on eye tracking, including:

[0094] Memory and processor; the memory is used to store computer-executable instructions, and the processor is used to execute computer-executable instructions to implement the VR product focus recognition method based on eye tracking as proposed in the above embodiment.

[0095] This embodiment also provides a storage medium having a computer program stored thereon, which, when executed by a processor, implements the VR product focus recognition method based on eye tracking as proposed in the above embodiment.

[0096] The storage medium proposed in this embodiment and the method for implementing VR product focus recognition based on eye tracking proposed in the above embodiment belong to the same inventive concept. For technical details not fully described in this embodiment, please refer to the above embodiment, and this embodiment has the same beneficial effects as the above embodiment.

[0097] Through the above description of the implementation methods, those skilled in the art can clearly understand that the present invention can be implemented with the help of software and necessary general-purpose hardware, and of course it can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art can be embodied in the form of a software product, which can be stored in a computer-readable storage medium, such as a computer's floppy disk, read-only memory (ROM), random access memory (RAM), flash memory (FLASH), hard disk or optical disk, etc., including a number of instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to execute the methods of various embodiments of the present invention.

[0098] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A VR product focus recognition method based on eye tracking, characterized in that: include: Use VR devices integrated with Tobii eye trackers to collect user gaze coordinates in real time; Inputting the collected gaze point coordinates into a convolutional neural network model, identifying the gaze target through the convolutional neural network model, and determining whether the gaze target is a product; If the gaze target is a product, the total time the user gazes at the product is counted, a product attention ranking result is obtained based on the total time, and the virtual shelf layout is adjusted according to the ranking result.

2. The method for VR product focus recognition based on eye tracking according to claim 1, characterized in that: Use VR devices with integrated Tobii eye trackers to collect user gaze coordinates in real time, including: Continuously collect the pupil center coordinates and corneal reflection point coordinates of the user's left and right eyes at a sampling frequency of 50Hz; Based on the pupil center coordinates and the corneal reflection point coordinates, the user's line of sight direction vector is calculated using the Purkinje imaging principle; Mapping the sight line direction vector to the three-dimensional space coordinate system of the VR scene, and determining the coordinates of the intersection of the sight line direction vector and the virtual product plane as the user's gaze point coordinates; The continuously collected gaze point coordinates are processed by Kalman filtering to eliminate the noise interference caused by eye tremors.

3. The method for VR product focus recognition based on eye tracking according to claim 2, characterized in that: The convolutional neural network model includes a feature extraction layer, a feature classification layer, and an attention mechanism layer; The feature extraction layer extracts image features of a preset pixel area around the gaze point, and the preset pixel area is dynamically adjusted according to the field of view and resolution of the VR device; Mapping the image features to a predefined target category space using a feature classification layer, where the target categories include product categories and UI interface categories; The image features extracted by the attention mechanism layer are then weighted to enhance the saliency of product features and suppress the interference of UI interface features. The convolutional neural network model is pre-trained using a transfer learning method, using the ImageNet dataset as the base training set, and then fine-tuned using product images and UI interface images in VR shopping scenarios.

4. The method for VR product focus recognition based on eye tracking according to claim 3, wherein: Identifying the gaze target by using the convolutional neural network model and determining whether the gaze target is a commodity includes: When the coordinates of the user's gaze point change, a local image block centered at the gaze point in the current frame of the VR scene image is captured; Inputting the local image block into a convolutional neural network model to obtain a classification probability distribution output by the model; If the classification probability of the commodity category exceeds the classification preset threshold, the gaze target is determined to be a commodity; If the classification results in consecutive frames are all UI interface categories, the current gaze point acquisition and recognition process is suspended until a change in the gaze point coordinates is detected.

5. The method for VR product focus recognition based on eye tracking according to claim 4, characterized in that: Obtaining a product attention ranking result based on the total duration includes: Create an attention record table for each product, recording the start time, end time, and duration of the product being looked at. Combine multiple attention events for the same product. If the time interval between two attention events is less than a preset interval threshold, the two attention durations are accumulated. Calculate the total gaze time of each product and sort them from high to low according to the total gaze time to obtain the product attention ranking results; The total gaze time is normalized and converted into an attention score of 0-100 to quantify the degree of attention received by the displayed products.

6. The method for VR product focus recognition based on eye tracking according to claim 5, characterized in that: Adjusting the virtual shelf layout according to the sorting result includes: Move the top 20% of products to the first visual area of ​​the virtual shelf; Dynamically adjust product display size based on product attention score; A virtual shelf layout optimization model is established, with product click-through rate and conversion rate as the objective function, and the product display location and display method are iteratively optimized through reinforcement learning algorithm.

7. The method for VR product focus recognition based on eye tracking according to claim 6, characterized in that: The total duration includes: If the total duration of a user's continuous gaze on an ad product exceeds the preset time threshold, the advertising billing mechanism is triggered.

8. A VR product focus recognition system based on eye tracking, characterized by: include, A data module is used to collect the user's gaze coordinates in real time using a VR device integrated with a Tobii eye tracker; A judgment module, configured to input the collected gaze point coordinates into a convolutional neural network model, identify the gaze target through the convolutional neural network model, and determine whether the gaze target is a commodity; The adjustment module is used to count the total time the user gazes at the product if the gaze target is a product, obtain a product attention ranking result based on the total time, and adjust the virtual shelf layout according to the ranking result.

9. An electronic device comprising: memory and processor; The memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions. When the computer-executable instructions are executed by the processor, the steps of the VR product focus recognition method based on eye tracking according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, implement the steps of the eye-tracking-based VR product focus recognition method according to any one of claims 1 to 7.

Images