A conference room function scene switching system based on an internet of things
Patent Information
- Application Number
- CN202211448257.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2022-11-18
- Publication Date
- 2026-08-21
- Estimated Expiration
- 2042-11-18
AI Technical Summary
[0002]会议室作为大多数公司内部工作讨论、团队建设等活动的重要场所，在工作以及日常活动中使用的时长也比较长；而大部分的会议室的背景均较为单一，在使用过程中往往不能显示出会议开展的重要性与风格；而对会议室的背景进行人工布展又往往需要耗费大量的人力物理；因此，需要一种根据会议室功能对会议室背景进行智能切换的场景切换系统；
本发明设置图片下载切换模块用于从后台服务器下载切换图片，并将会议室背景切换为对应图片；设置风格配置模块预先收集实验室常用功能对应的会议风格；设置图片风格匹配模块训练神经网络模型自动为图片计算每类风格的概率；设置会议信息输入模块远程向后台服务器提供待使用会议室编号以及会议室功能信息；设置会议风格提取模块训练神经网络模型为每个功能信息计算每类风格的概率；设置会议室图片选择模块根据图片风格概率以及功能信息风格概率获得风格最匹配的背景图片；解决了会议室内背景智能切换问题。
Smart Images
Figure CN115689874B_ABST
Abstract
Description
Technical Field
[0001] This invention belongs to the field of conference room function switching and involves Internet of Things (IoT) technology. Specifically, it is an IoT-based conference room function scene switching system. Background Technology
[0002] Meeting rooms are important venues for internal work discussions, team building activities, and other activities in most companies, and they are used for a long time in work and daily activities. However, most meeting room backgrounds are relatively simple and often fail to reflect the importance and style of the meeting. Manually setting up meeting room backgrounds often requires a lot of manpower and resources. Therefore, there is a need for a scene switching system that can intelligently switch meeting room backgrounds according to the function of the meeting room. To address this, an IoT-based conference room function scene switching system is proposed. Summary of the Invention
[0003] This invention aims to at least solve one of the technical problems existing in the prior art. To this end, this invention proposes an IoT-based conference room function scene switching system. This system includes an image download and switching module for downloading switching images from a backend server and changing the conference room background to the corresponding image; a style configuration module for pre-collecting meeting styles corresponding to commonly used laboratory functions; an image style matching module for training a neural network model to automatically calculate the probability of each style for each image; a meeting information input module for remotely providing the conference room number and function information to the backend server; a meeting style extraction module for training a neural network model to calculate the probability of each style for each function; and a conference room image selection module for obtaining the most matching background image based on the image style probability and the function information style probability. This solves the problem of intelligent background switching in conference rooms.
[0004] To achieve the above objectives, an Internet of Things-based conference room function scene switching system is proposed according to an embodiment of the first aspect of the present invention, including an image download and switching module, a backend server, a style configuration module, an image style matching module, a conference information input module, a conference style extraction module, and a conference room image selection module; The image download and switching module is mainly used to download images of different styles from the backend server and switch the meeting style of the meeting room by modifying the image of the background panel in the meeting room. The image download and switching module includes an image download device and a conference room backdrop. The image download device downloads a background image from a backend server via a wireless network and sends the downloaded background image to the conference room backdrop via electrical means. The conference room backdrop can be an LCD screen. After receiving the background image, the conference room backdrop adapts the image size to the size of the LCD screen and displays it on the LCD screen. The style configuration module is mainly used to pre-configure the conference room style according to the conference content. The style configuration module pre-determines meeting styles by manually collecting common meeting content in meeting rooms based on practical experience; the meeting style is the functional embodiment of the meeting, and the style configuration module saves the collected meeting styles in the form of a set on the backend server; the style set is marked as S; each style category is marked as s; The image style matching module is mainly used to perform style labeling for each image using neural network technology. The image style matching module performs style labeling for each image, including the following steps: Step S1: The image style matching module collects several background images corresponding to each style s in the style set S in advance; and tags each background image with the corresponding style tag; Step S2: The image style matching module takes the background image as input and feeds it into the CNN neural network model; the CNN neural network model outputs the predicted style of the background image; the accuracy between the predicted style and the actual style of the background image is used as the target of the neural network; the CNN neural network model is trained; the parameters of the CNN neural network model are set and adjusted according to practical experience. Step S3: The image style matching module stops training the CNN neural network model when the accuracy reaches 98%; the trained CNN neural network model is marked as M; Step S4: The image style matching module collects all background images used for actual switching in the conference room; Step S5: The image style matching module takes each actual background image as input and inputs it into the CNN neural network model M to obtain the probability that each background image belongs to each style; the set of each actual background image is labeled as P; then each background image is labeled as p; further, the probability that background image p belongs to style s is labeled as Gps; The image style matching module stores the style probability GPS of all background images in the background server; The meeting information input module is mainly used to input the functions that need to be carried out in the meeting room before it is used. The meeting information input module includes a remote input device; this remote input device is connected to the backend server via a wireless network, and staff can input meeting room reservation information into the remote input device. The meeting room reservation information includes, but is not limited to, the meeting room number and meeting room reservation function information; the meeting information input module sends the meeting room reservation information entered by the user to the meeting room image selection module. The meeting style extraction module is mainly used to analyze the style required for the meeting room based on the meeting room reservation information input by the user. The meeting style extraction module analyzes the style that the meeting room needs to use, including the following steps: Step P1: The meeting style extraction module pre-collects textual information about the functions of several meeting rooms; and assigns a corresponding meeting style tag to each function information by manual annotation; the meeting style tag is a style in the style tag set S; Step P2: The meeting style extraction module segments the functional text information into words and converts them into word vectors. The word vectors are then input into the RNN neural network model. The RNN neural network model outputs the predicted meeting style. The accuracy between the predicted meeting style and the actual style of the functional text information is used as the target of the RNN neural network. The RNN neural network model is trained. The parameters of the RNN neural network model are set and adjusted according to practical experience. Step P3: The meeting style extraction module stops training the RNN neural network model when the accuracy reaches 98%; the trained RNN neural network model is labeled as R; The meeting style extraction module sends the trained RNN neural network model R to the meeting room image selection module; The meeting room image selection module is mainly used to use an RNN neural network model to determine the style corresponding to the actual function required by the meeting room, and select images with similar styles to display based on the style of the function information. The image selection module for the conference room includes the following steps: Step Q1: The meeting room image selection module segments the reserved meeting room function information in the meeting room reservation information sent by the user and converts it into word vector form; inputs the word vectors into the RNN neural network model R to obtain the probability of each style; sorts the probability of the function information belonging to style s in descending order of probability; and labels the sorted style set as Sp; each style in the style set Sp is labeled as n; the probability of each style n is labeled as Yn; where n=1,2,…,|S|; Step Q2: Set i level coefficients based on practical experience; label each level coefficient as Dn; and satisfy 0 <D1<D2<D3<…<Di; Step Q3: Calculate the style similarity coefficient Fp between each image p in the image set P and the meeting room function information; where the formula for calculating the style similarity coefficient Fp is... ; Step Q4: Find the image with the smallest style similarity coefficient (Fp) from all images and use it as the switching image for the meeting room; The conference room image will be changed and sent to the image download and switching module. The image download and switching module will then download the image and change the conference room background from the backend server based on the changed image name.
[0005] Compared with the prior art, the beneficial effects of the present invention are: This invention includes an image download and switching module for downloading and switching images from a backend server and changing the meeting room background to the corresponding image; a style configuration module for pre-collecting meeting styles corresponding to commonly used laboratory functions; an image style matching module for training a neural network model to automatically calculate the probability of each style for each image; a meeting information input module for remotely providing the meeting room number and function information to the backend server; a meeting style extraction module for training a neural network model to calculate the probability of each style for each function; and a meeting room image selection module for obtaining the most matching background image based on the image style probability and the function information style probability. This invention solves the problem of intelligent background switching in meeting rooms. Attached Figure Description
[0006] Figure 1 This is a schematic diagram of the present invention. Detailed Implementation
[0007] The technical solution of the present invention will be clearly and completely described below with reference to the embodiments. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.
[0008] like Figure 1 As shown, an IoT-based conference room function scene switching system includes an image download and switching module, a backend server, a style configuration module, an image style matching module, a conference information input module, a conference style extraction module, and a conference room image selection module. The image download and switching module is mainly used to download images of different styles from the backend server and switch the meeting style of the meeting room by modifying the image of the background panel in the meeting room. In a preferred embodiment, the image download switching module includes an image download device and a conference room background board; wherein, the image download device downloads a background image from a backend server via a wireless network and sends the downloaded background image to the conference room background board via an electrical means; the conference room background board can be an LCD screen; after receiving the background image, the conference room background board adapts the size of the background image to the size of the LCD screen and displays it on the LCD screen; The style configuration module is mainly used to pre-configure the conference room style according to the conference content. In a preferred embodiment, the style configuration module pre-determines the meeting style by manually collecting common meeting content in the meeting room based on practical experience; the meeting style reflects the function of the meeting, such as: teaching meeting, work meeting, team building activity, etc.; it is understood that the required meeting room background image is different under different meeting functions; the style configuration module saves the collected meeting styles in the form of a set in the backend server; the style set is marked as S; each style category is marked as s; The image style matching module is mainly used to perform style labeling for each image using neural network technology. Each background image has its own stylistic characteristics, but since the number of background images is often quite large, manual style labeling would require a huge amount of manpower. Therefore, we are considering using neural network technology to label the style of images to reduce manpower costs. The image style matching module performs style labeling for each image, including the following steps: Step S1: The image style matching module collects several background images corresponding to each style s in the style set S in advance; and tags each background image with the corresponding style tag; Step S2: The image style matching module takes the background image as input and feeds it into the CNN neural network model; the CNN neural network model outputs the predicted style of the background image; the accuracy between the predicted style and the actual style of the background image is used as the target of the neural network; the CNN neural network model is trained; the parameters of the CNN neural network model are set and adjusted according to practical experience. Step S3: The image style matching module stops training the CNN neural network model when the accuracy reaches 98%; the trained CNN neural network model is marked as M; Step S4: The image style matching module collects all background images used for actual switching in the conference room; Step S5: The image style matching module takes each actual background image as input and inputs it into the CNN neural network model M to obtain the probability that each background image belongs to each style; the set of each actual background image is labeled as P; then each background image is labeled as p; further, the probability that background image p belongs to style s is labeled as Gps; The image style matching module stores the style probability GPS of all background images in the background server; The meeting information input module is mainly used to input the functions that need to be carried out in the meeting room before it is used. Understandably, in daily work, meeting rooms often need to be booked in advance, and when booking a meeting room, the reason for booking and the purpose of booking need to be filled in. In a preferred embodiment, the meeting information input module includes a remote input device; the remote input device is connected to the backend server via a wireless network, and staff input meeting room reservation information into the remote input device, the meeting room reservation information including but not limited to the meeting room number and meeting room reservation function information; the meeting information input module sends the meeting room reservation information input by the user to the meeting room image selection module; The meeting style extraction module is mainly used to analyze the style required for the meeting room based on the meeting room reservation information input by the user. In a preferred embodiment, the meeting style extraction module analyzes the style required for the meeting room by including the following steps: Step P1: The meeting style extraction module pre-collects textual information about the functions of several meeting rooms; and assigns a corresponding meeting style tag to each function information by manual annotation; the meeting style tag is a style in the style tag set S; Step P2: The meeting style extraction module segments the functional text information into words and converts them into word vectors. The word vectors are then input into the RNN neural network model. The RNN neural network model outputs the predicted meeting style. The accuracy between the predicted meeting style and the actual style of the functional text information is used as the target of the RNN neural network. The RNN neural network model is trained. The parameters of the RNN neural network model are set and adjusted according to practical experience. Step P3: The meeting style extraction module stops training the RNN neural network model when the accuracy reaches 98%; the trained RNN neural network model is labeled as R; The meeting style extraction module sends the trained RNN neural network model R to the meeting room image selection module; The meeting room image selection module is mainly used to use an RNN neural network model to determine the style corresponding to the actual function required by the meeting room, and select images with similar styles to display based on the style of the function information. In a preferred embodiment, the conference room image selection module selects images by including the following steps: Step Q1: The meeting room image selection module segments the reserved meeting room function information in the meeting room reservation information sent by the user and converts it into word vector form; inputs the word vectors into the RNN neural network model R to obtain the probability of each style; sorts the probability of the function information belonging to style s in descending order of probability; and labels the sorted style set as Sp; each style in the style set Sp is labeled as n; the probability of each style n is labeled as Yn; where n=1,2,…,|S|; Step Q2: Set i level coefficients based on practical experience; label each level coefficient as Dn; and satisfy 0 <D1<D2<D3<…<Di; Step Q3: Calculate the style similarity coefficient Fp between each image p in the image set P and the meeting room function information; where the formula for calculating the style similarity coefficient Fp is... ; Step Q4: Find the image with the smallest style similarity coefficient (Fp) from all images and use it as the switching image for the meeting room; The conference room image will be changed and sent to the image download and switching module. The image download and switching module will then download the image and change the conference room background from the backend server based on the changed image name.
[0009] The above embodiments are only used to illustrate the technical methods of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical methods of the present invention without departing from the spirit and scope of the technical methods of the present invention.
Claims
1. A conference room function scene switching system based on the Internet of Things, characterized in that, It includes an image download and switching module, a backend server, a style configuration module, an image style matching module, a meeting information input module, a meeting style extraction module, and a meeting room image selection module; The image download and switching module is used to download images of different styles from the backend server and switch the meeting style of the meeting room by modifying the image of the background panel in the meeting room. The style configuration module is used to pre-configure the meeting room style according to the meeting content; the style configuration module saves the collected meeting styles in the form of a set on the backend server; the style set is marked as S; each style category is marked as s; The image style matching module is used to perform style labeling for each image using neural network technology; the image style matching module stores the background image style probability GPS in the background server; The meeting information input module is used to input the functions to be carried out in the meeting room before its use; the meeting information input module sends the meeting room reservation information entered by the user to the meeting room image selection module; The meeting style extraction module uses neural network technology to analyze the required style of the meeting room based on the user-input meeting room reservation information; the meeting style extraction module sends the trained RNN neural network model R to the meeting room image selection module; The meeting room image selection module is used to use a neural network model to determine the style corresponding to the function required by the meeting room, and select images with similar styles to display based on the style of the function information. The image style matching module performs style labeling for each image, including the following steps: Step S1: The image style matching module collects several background images corresponding to each style s in the style set S in advance; and assigns a corresponding style label to each background image; Step S2: The image style matching module takes the background image as input and inputs it into the CNN neural network model; the CNN neural network model takes the predicted style of the background image as output; and trains the CNN neural network model; Step S3: The image style matching module stops training the CNN neural network model when the accuracy reaches 98%; the trained CNN neural network model is marked as M; Step S4: The image style matching module collects all background images used for actual switching in the conference room; Step S5: The image style matching module takes each actual background image as input and inputs it into the CNN neural network model M to obtain the probability that each background image belongs to each style; the set of each actual background image is labeled as P; then each background image is labeled as p; the probability that background image p belongs to style s is labeled as Gps; The meeting style extraction module analyzes the style that the meeting room needs to use, including the following steps: Step P1: The meeting style extraction module pre-collects textual information about the functions of several meeting rooms; and assigns a corresponding meeting style tag to each function information by manual annotation. Step P2: The meeting style extraction module segments the functional text information into word vectors and inputs the word vectors into the RNN neural network model; the RNN neural network model outputs the predicted meeting style; the RNN neural network model is then trained. Step P3: The meeting style extraction module stops training the RNN neural network model when the accuracy reaches 98%; the trained RNN neural network model is labeled as R; The image selection module for the conference room includes the following steps: Step Q1: The meeting room image selection module segments the reserved meeting room function information in the meeting room reservation information sent by the user and converts it into word vector form; inputs the word vectors into the RNN neural network model R to obtain the probability of each style; sorts the probability of the function information belonging to style s in descending order of probability; and labels the sorted style set as Sp; each style in the style set Sp is labeled as n; the probability of each style n is labeled as Yn; where n=1,2,…,|S|; Step Q2: Set i level coefficients based on practical experience; label each level coefficient as Di; and satisfy 0 <D1<D2<D3<…<Di; Step Q3: Calculate the style similarity coefficient Fp between each image p in the image set P and the meeting room function information; where the formula for calculating the style similarity coefficient Fp is... Fp=∑ n Di (Yn-Gpn) 2 ； Step Q4: Find the image with the smallest style similarity coefficient Fp from all images and use it as the switching image for the meeting room.
2. The IoT-based conference room function scene switching system according to claim 1, characterized in that, The image download and switching module includes an image download device and a conference room backdrop. The image download device downloads a background image from a backend server via a wireless network and sends the downloaded background image to the conference room backdrop via electrical means. After receiving the background image, the conference room backdrop adapts the image size to the size of the LCD screen and displays it on the LCD screen.
3. The IoT-based conference room function scene switching system according to claim 2, characterized in that, The style configuration module pre-determines the meeting style based on practical experience and by manually collecting common meeting content in the conference room.
4. The IoT-based conference room function scene switching system according to claim 3, characterized in that, The meeting information input module includes a remote input device; the remote input device is connected to the back-end server via a wireless network, and staff can input meeting room reservation information into the remote input device.
5. The IoT-based conference room function scene switching system according to claim 4, characterized in that, The conference room image will be changed and sent to the image download and switching module. The image download and switching module will then download the image and change the conference room background from the backend server based on the changed image name.
Citation Information
Patent Citations
Decoration style recommendation system and method based on artificial intelligence
CN114996498A