TW202110211A

TW202110211A - Voice advertisement system capable of being broadcasted to specific persons and its implementation method capable of being broadcasted audio advertising to a specific person without being received by other non-labeled customers

Info

Publication number: TW202110211A
Application number: TW108130920A
Authority: TW
Inventors: 薛樂山
Original assignee: 大陸商南京深視光點科技有限公司
Priority date: 2019-08-28
Filing date: 2019-08-28
Publication date: 2021-03-01

Abstract

A voice advertisement system that can be broadcasted to specific persons and its implementation method mainly has a three-dimensional image sensing module arranged in indoor and outdoor environments, which will recognize the specific biological characteristics of the passing consumers, such as gender, age, height, weight, etc. After further tracking the location of the specific consumer in the three-dimensional space, a directional sound source module is used to project a specific directional advertising audio to the specific consumer according to the characteristics. Other consumers who are not labeled in the limited space but can be sensed by the three-dimensional image sensing module will not receive the advertising audio. The invention has the effect that different advertising audio can be broadcasted to different population groups within the same space, and the different advertising audio do not affect each other.

Description

可投放於特定人員的語音廣告系統及其實施方法Voice advertisement system capable of being placed on specific personnel and its implementation method

一種可投放於特定人員的語音廣告系統及其實施方法，尤指一種可在限定範圍空間內持續追蹤一特定人員，並發送一其他人聽不見的廣告音訊給該特定人員的語音廣告系統。A voice advertisement system that can be placed on a specific person and an implementation method thereof, especially a voice advertisement system that can continuously track a specific person in a limited space and send an advertisement message that other people can't hear to the specific person.

一般無論在百貨公司或是商場等商業場所，推播廣告訊息的手法很常見，例如在牆上設置靜態的海報或動態的電子看板，或是以語音廣播的方式在開放空間進行循環式的播放等，請參閱「第1圖」，圖中所示為習知商場的語音廣告推播示意圖，如圖中所示的一商場10，其主要係透過播放裝置對商場10的一開放空間進行廣播式的語音廣告播放，以供路過開放空間的消費者聽到商家推播的語音廣告，但是開放空間中的語音廣告可能不只有一則，多則語音廣告同時在開放空間中進行推播時，亦容易造成資訊混亂的狀況，且上述的廣告訊息推播方式都無法精準的掌握客群屬性，只能以亂槍打鳥的方式進行廣告推播，以商家的立場來說是非常沒有效益的；請參照中華民國專利第TW201814644「基於虛擬圍欄的信息處理方法、用戶端及伺服器」，其主要利用一個虛擬圍欄(Geo-fencing)的圍出一個虛擬地理邊界，當消費者進入虛擬圍欄區域時，手機可以自動接受廣告訊息，藉此技術可對特定區域內的路人推播適當的廣告訊息，另有其他專利以供參考：（1）中華民國專利第TWI554963「無線信號移動信息傳送方法和系統」；（2）中國專利第CN108269133A「結合人體識別和語音辨識的智慧廣告推送方法及終端」；（3）美國專利第US20180018508A1「COMPUTER VISION SYSTEMS」。Generally, whether in department stores or shopping malls and other commercial places, the methods of pushing and broadcasting advertising messages are very common, such as setting up static posters or dynamic electronic billboards on the walls, or playing cyclically in open spaces by means of voice broadcasts. Please refer to "Figure 1". The figure shows a schematic diagram of the voice advertisement push broadcast of the conventional shopping mall. As shown in the figure, a shopping mall 10 is mainly used to broadcast an open space of the shopping mall 10 through a broadcasting device. Voice advertisements are played in a manner so that consumers passing by the open space can hear the voice advertisements promoted by the business. However, there may be more than one voice advertisement in the open space. When multiple voice advertisements are promoted in the open space at the same time, it is also easy This has caused information confusion, and none of the above advertising message push methods can accurately grasp the attributes of the customer group. The advertising push can only be carried out in a random manner, which is very inefficient from the standpoint of the business; please; Refer to the Republic of China Patent No. TW201814644 "Information Processing Method, Client and Server Based on Virtual Fence", which mainly uses a virtual fence (Geo-fencing) to enclose a virtual geographic boundary. When consumers enter the virtual fence area, Mobile phones can automatically receive advertising messages, and this technology can push appropriate advertising messages to passers-by in a specific area. There are other patents for reference: (1) The Republic of China Patent No. TWI554963 "Wireless Signal Mobile Information Transmission Method and System"; (2) Chinese Patent No. CN108269133A "Smart advertisement push method and terminal combining human body recognition and voice recognition"; (3) US Patent No. US20180018508A1 "COMPUTER VISION SYSTEMS".

現有的廣告推播方式大多是以被動推播為主，且主要的缺點是無法有效辨識客群屬性推播適當的廣告訊息，未分類的廣告訊息因其被廣泛播放，對大部分的消費者而言是不具價值的訊息，因此大幅減少商家的成交機會。Most of the existing advertising push methods are mainly passive push, and the main disadvantage is that they cannot effectively identify the attributes of the customer group to push appropriate advertising messages. Uncategorized advertising messages are widely broadcast, which is very important to most consumers. It is a message of no value, so it greatly reduces the merchant’s chances of transaction.

有鑑於上述的問題，本發明人係依據多年來開發廣告推播系統的經驗，針對開放式空間推播廣告訊息的方法及系統進行研究及分析，期能研發出解決上述問題的方法及系統；緣此，本發明之主要目的在於提供一種可辨識開放空間範圍內的消費者之生物特徵，藉以推播適合該消費者屬性的廣告訊息，且系統所推播的廣告訊息可在不影響他人的狀況下進行投放的語音廣告系統。In view of the above-mentioned problems, the inventor of the present invention has conducted research and analysis on methods and systems for the promotion of advertising messages in open spaces based on years of experience in developing advertising promotion systems, and hopes to develop methods and systems to solve the above-mentioned problems; For this reason, the main purpose of the present invention is to provide a biometric feature that can identify consumers in the open space, so as to promote advertising messages suitable for the attributes of the consumers, and the advertising messages promoted by the system can be used without affecting others’ A voice advertising system for delivery under conditions.

為達上述的目的，本發明主要透過三維影像感測模組辨識出消費者其特定生物特徵，例如性別、年齡、身高等，以確定該消費者的屬性，再進一步追蹤標定該消費者後，搭配指向性音源模組投射出特定的廣告音訊給該特定消費者，而空間範圍內的其他非被標定之消費者，則不會收到廣告音訊，使本發明具有在同一空間範圍內可針對不同族群投放不同廣告音訊，且不同的廣告音訊又不相互影響的功效。In order to achieve the above-mentioned purpose, the present invention mainly uses the three-dimensional image sensing module to identify the consumer's specific biological characteristics, such as gender, age, height, etc., to determine the consumer's attributes, and then further track and calibrate the consumer, The directional sound source module is used to project a specific advertising message to the specific consumer, and other non-marked consumers in the space will not receive the advertising message, so that the present invention can be targeted within the same space. Different groups of people put different advertising messages, and different advertising messages do not affect each other.

為使貴審查委員得以清楚了解本發明之目的、技術特徵及其實施後之功效，茲以下列說明搭配圖示進行說明，敬請參閱。In order to enable your reviewer to have a clear understanding of the purpose, technical features and effects of the present invention after implementation, the following descriptions and illustrations are used for explanation, please refer to them.

請參閱「第2圖」，圖中所示為本發明之系統架構圖（一），如圖中所示的可投放於特定人員的語音廣告系統20，其主要係由一三維影像感測模組201、一指向性音源模組202以及一管理伺服器203所組構而成，且管理伺服器203分別與三維影像感測模組201以及指向性音源模組202形成資訊連線；其中，所述的三維影像感測模組201具有一辨識單元2011以及一追蹤單元2012，且所述的三維影像感測模組201主要為一深度攝影機(Depth Camera)，其可測量出每一個影像點和攝影機之間的Z軸距離，因此所儲存的是三維的空間影像資訊，而三維影像感測模組201可在擷取影像畫面後產生一影像資訊A，所述的影像資訊A可透過辨識單元2011，對經過的人進行人體生物特徵的辨識，例如性別、年齡、身高、體重等，並可進一步透過追蹤單元2012來追蹤影像資訊A中移動的人體；指向性音源模組202可發出指向性廣告音訊，所述的指向性廣告音訊即使在空間中傳播，人耳也無法查覺指向性廣告音訊的存在，意即在指向性音源模組202所發出的音波束區域內，其所發出的廣告音訊是人耳聽得見的一般音波，而在音波束區域以外則是人耳聽不見的超音波；管理伺服器203分別與三維影像感測模組201以及指向性音源模組202形成資訊連結，管理伺服器203具有可供以儲存或提供資訊存取的一資料庫2031，可儲存由一廣告商或商家所提供之複數廣告等相關的廣告音訊資料N；再請搭配參閱「第3圖」，圖中所示為本發明之系統架構圖（二），承上所述，管理伺服器203具有一族群分類單元2032，所述的族群分類單元2032可依據消費者的生理特徵做分類，例如圖中所示的族群分類單元2032，其詳述了男女性別搭配各階段年齡層所分類出各種階段的族群，例如族群A為女性20歲以下，適合洗髮乳或手機等年輕族群的廣告，而族群I則適合汽車等中年族群的廣告。Please refer to "Figure 2". The figure shows the system architecture diagram (1) of the present invention. The voice advertisement system 20 shown in the figure that can be placed on a specific person is mainly composed of a three-dimensional image sensing model. The group 201, a directional sound source module 202, and a management server 203 are assembled, and the management server 203 respectively forms an information connection with the 3D image sensing module 201 and the directional sound source module 202; wherein, The 3D image sensing module 201 has an identification unit 2011 and a tracking unit 2012, and the 3D image sensing module 201 is mainly a depth camera, which can measure each image point The Z-axis distance between the camera and the camera, so the three-dimensional spatial image information is stored, and the three-dimensional image sensing module 201 can generate an image information A after capturing the image frame. The image information A can be identified by The unit 2011 identifies the human biological characteristics of the passing people, such as gender, age, height, weight, etc., and can further track the moving human body in the image information A through the tracking unit 2012; the directional sound source module 202 can emit pointing The directional advertisement information, even if the directional advertisement information is spread in space, the human ear cannot detect the existence of the directional advertisement information, which means that the sound beam area emitted by the directional sound source module 202 is emitted by the human ear. The advertising audio of is the normal sound waves that can be heard by the human ear, while outside the sound beam area is the ultrasonic waves that cannot be heard by the human ear; the management server 203 is respectively formed with the 3D image sensing module 201 and the directional sound source module 202 Information link. The management server 203 has a database 2031 that can be used to store or provide information access. It can store multiple advertisements and other related advertising information data N provided by an advertiser or business; Figure 3" shows the system architecture diagram (2) of the present invention. As mentioned above, the management server 203 has an ethnic group classification unit 2032, which can be based on the physiological characteristics of consumers. Classification, such as the ethnic classification unit 2032 shown in the figure, which details the various stages of the ethnic groups classified by the age groups of the genders of men and women. For example, ethnic group A is female under 20 years old, suitable for young people such as shampoo or mobile phones Advertisements, and group I is suitable for middle-aged groups such as automobiles.

請參閱「第4圖」及「第5圖」，圖中所示為本發明之實施示意圖（一）、（二），請搭配參閱「第1圖」～「第3圖」，本發明於實施時，係預先以三維影像感測模組201對一開放空間P內的數名走動的消費者進行影像擷取作業，而三維影像感測模組201的數量可為一組以上，並進一步將所擷取的影像資訊，透過辨識單元2011進行人體生物特徵的辨識，例如性別、年齡、身高、體重等，並在完成辨識後產生多筆辨識資訊F，所述的多筆辨識資訊F即代表多名消費者的特徵資訊，且數筆辨識資訊F可分別被傳送至管理伺服器203的族群分類單元2032進行比對分類，以產生數筆相對應辨識資訊F的建議推播資訊G，管理伺服器203再進一步將數筆建議推播資訊G與資料庫2031中的廣告音訊資料N配對，以挑選出適合建議推播資訊G的廣告音訊資料；再請搭配參閱「第6圖」，圖中所示為本發明之實施示意圖（三），承「第4圖」及「第5圖」所述，三維影像感測模組201的追蹤單元2012可追蹤畫面中多名消費者的移動軌跡，因此，管理伺服器203可搭配追蹤單元2012所取得的位置資訊，進一步驅動指向性音源模組202，使其持續對其中一名被標定的消費者，進行廣告音訊資料N的投放，而該名被標定的消費者因位於指向性音源模組202所發出的音波束區域內，因此可聽到指向性音源模組202所投放的廣告音訊資料N，而未被標定的消費者由於在指向性音源模組202所發出的音波束區域外，因此無法聽到指向性音源模組202所投放的廣告音訊資料N；藉此，系統可針對消費者屬性推播適合的廣告音訊資料N，有效增加成交機率，且系統推播的廣告音訊資料N可在不影響他人的狀況下進行投放，可避免賣場的消費環境中充斥著多則廣告音訊到處播放的狀況產生；再者，本發明所述的指向性音源模組202其數量非為限定一組，本發明為便於說明僅以一組做為實施說明，非為限定本發明指向性音源模組202的數量，本發明所述的指向性音源模組202可為多組實施，以同時對多名消費者投放適合其屬性的廣告音訊資料N。Please refer to "Figure 4" and "Figure 5". The figures show the schematic diagrams (1) and (2) of the implementation of the present invention. Please refer to "Figure 1" to "Figure 3". In implementation, the three-dimensional image sensing module 201 is used to perform image capturing operations on several walking consumers in an open space P in advance, and the number of the three-dimensional image sensing module 201 can be more than one set, and further The captured image information is identified through the identification unit 2011 to identify human biological characteristics, such as gender, age, height, weight, etc., and after the identification is completed, multiple pieces of identification information F are generated. The multiple pieces of identification information F are Characteristic information representing multiple consumers, and several pieces of identification information F can be respectively sent to the ethnic group classification unit 2032 of the management server 203 for comparison and classification to generate several pieces of recommended push information G corresponding to the identification information F. The management server 203 further matches several pieces of recommended push information G with the advertisement audio data N in the database 2031 to select advertisement audio data suitable for the recommended push information G; please refer to "Figure 6" for the combination, Shown in the figure is a schematic diagram (3) of the implementation of the present invention. As described in "Figure 4" and "Figure 5", the tracking unit 2012 of the 3D image sensor module 201 can track the movement of multiple consumers on the screen Therefore, the management server 203 can cooperate with the location information obtained by the tracking unit 2012 to further drive the directional audio source module 202 to continuously deliver advertising audio data N to one of the identified consumers, and Since the calibrated consumer is located in the sound beam area emitted by the directional sound source module 202, he can hear the advertising audio data N delivered by the directional sound source module 202, while the uncalibrated consumer is pointing at The sound beam area emitted by the sexual sound source module 202 cannot be heard. Therefore, the advertising audio data N delivered by the directional sound source module 202 cannot be heard; thereby, the system can push and broadcast suitable advertising audio data N according to the attributes of consumers, effectively increasing Probability of the transaction, and the advertising audio data N pushed by the system can be placed without affecting others, which can avoid the situation that the consumer environment of the store is flooded with multiple advertising audios being played everywhere; furthermore, the invention described in the present invention The number of directional sound source modules 202 is not limited to one set. For the convenience of description, the present invention only uses one set as an implementation description. It is not meant to limit the number of directional sound source modules 202 of the present invention. The module 202 can be implemented in multiple groups to simultaneously deliver advertisement audio data N suitable for their attributes to multiple consumers.

請參閱「第7圖」，圖中所示為本發明之實施步驟流程示意圖，請搭配參閱「第1圖」～「第6圖」，本發明實施步驟如下: (1) 影像擷取步驟S1：系統利用三維影像感測模組201對一開放空間P內的數名消費者進行影像擷取作業，並將所擷取的影像資訊利用辨識單元2011進行人體生物特徵的辨識，例如性別、年齡、身高、體重等，並在完成辨識後產生多筆辨識資訊F； (2) 消費者屬性比對步驟S2：數筆辨識資訊F可分別被傳送至管理伺服器203的族群分類單元2032進行比對分類，以產生數筆相對應辨識資訊F的建議推播資訊G； (3) 廣告音訊配對步驟S3：管理伺服器203進一步將數筆建議推播資訊G與資料庫2031中儲存的數筆廣告音訊資料進行配對，以挑選出符合建議推播資訊G的廣告音訊資料； (4) 投放廣告音訊步驟S4：管理伺服器203可搭配追蹤單元2012所取得的位置資訊，進一步驅動指向性音源模組202持續對其中一名被標定的消費者進行廣告音訊資料N的投放。Please refer to "Figure 7", which shows a schematic diagram of the flow of implementation steps of the present invention. Please refer to "Figure 1" to "Figure 6" together. The implementation steps of the present invention are as follows: (1) Image capturing step S1: The system uses the three-dimensional image sensing module 201 to perform image capturing operations on several consumers in an open space P, and uses the captured image information to perform human biological identification using the identification unit 2011 Identification of features, such as gender, age, height, weight, etc., and multiple pieces of identification information F are generated after the identification is completed; (2) Consumer attribute comparison step S2: Several pieces of identification information F can be respectively sent to the ethnic classification unit 2032 of the management server 203 for comparison and classification, so as to generate several pieces of recommended push information G corresponding to the identification information F ； (3) Advertisement audio matching step S3: The management server 203 further matches several pieces of recommended information G with several pieces of advertisement audio data stored in the database 2031 to select advertisement audio data that meet the recommended information G ； (4) Placing advertising audio step S4: The management server 203 can cooperate with the location information obtained by the tracking unit 2012 to further drive the directional audio source module 202 to continuously deliver advertising audio data N to one of the targeted consumers.

請參閱「第8圖」，圖中所示為本發明之另一實施例（一），由上述可知，被標定的消費者所接收到的廣告音訊資料N，主要係由系統經過客群屬性分析後所進行播放，然而，這只是系統經過大數據統計之後的數據，未必每個消費者都能接受被推播的廣告音訊資料N，因此，為避免消費者有不好的廣告接收體驗，本發明係進一步在三維影像感測模組201中設有一情緒特徵辨識單元2013，並在管理伺服器203的資料庫2031中，預先儲存有數筆臉部特徵資訊H，所述的臉部特徵資訊H主要為一臉部表情辨識模型資訊，更詳細地說明，情緒特徵辨識單元2013主要透過一神經網路(Neural Network)模型訓練出臉部特徵辨識模型，以辨識複數臉部影像中的臉部特徵是否與資料庫2031中儲存的數筆臉部特徵資訊H吻合，其中上述臉部特徵資訊H包括性別、年齡以及情緒，且情緒可包括正向情緒及負向情緒，其中正向情緒係為興奮、快樂、愉悅、輕鬆、寧靜及冷靜其中之一，負向情緒係為苦惱、生氣、緊張、悲傷、無聊及困倦其中之一，而神經網路可以用C++、Visual Basic或其他適用於此目的的物件導向程式語言來實現，當消費者接收系統所推播的廣告音訊資料N後，系統即可以三維影像感測模組201擷取消費者的臉部影像，並進一步透過情緒特徵辨識單元2013辨識出該消費者目前的臉部情緒，並與資料庫2031中儲存的數筆臉部特徵資訊H進行比對分析，以產生一情緒資訊，若情緒資訊比對結果為正向情緒，則系統持續推播廣告音訊資料N，若情緒資訊比對結果為負向情緒，表示消費者對目前所推播的廣告音訊資料N感到厭惡或不滿意，此時，三維影像感測模組201會將情緒資訊傳送至管理伺服器203，藉由管理伺服器203驅使指向性音源模組202停止繼續推播廣告音訊資料N，或是驅動指向性音源模組202推播其他適合的廣告音訊資料N，以避免造成消費者不好的消費體驗。Please refer to "Figure 8". The figure shows another embodiment (1) of the present invention. From the above, it can be seen that the advertising audio data N received by the marked consumer is mainly caused by the system through the customer group attribute It is played after analysis. However, this is only the data collected by the system after big data statistics. Not every consumer can accept the pushed advertising audio data N. Therefore, in order to prevent consumers from having a bad advertising receiving experience, In the present invention, an emotional feature recognition unit 2013 is further provided in the three-dimensional image sensing module 201, and a number of facial feature information H is pre-stored in the database 2031 of the management server 203, the facial feature information H is mainly a facial expression recognition model information. In more detail, the emotion feature recognition unit 2013 mainly trains a facial feature recognition model through a Neural Network model to recognize faces in plural facial images Whether the feature matches the facial feature information H stored in the database 2031, where the facial feature information H includes gender, age, and emotion, and the emotion can include positive emotion and negative emotion, where the positive emotion system is One of excitement, happiness, pleasure, relaxation, tranquility, and calmness. Negative emotions are one of distress, anger, tension, sadness, boredom and sleepiness. The neural network can be used in C++, Visual Basic or other applications. The objective object-oriented programming language is realized. After the consumer receives the advertising audio data N pushed by the system, the system can capture the consumer’s facial image by the three-dimensional image sensing module 201, and further use the emotional feature recognition unit 2013 identifies the consumer’s current facial emotions and compares and analyzes it with several facial feature information H stored in the database 2031 to generate emotion information. If the emotion information comparison result is a positive emotion, then The system continues to push advertising audio data N. If the comparison result of emotional information is negative emotion, it means that consumers are disgusted or dissatisfied with the advertising audio data N currently pushed. At this time, the 3D image sensing module 201 will Send emotional information to the management server 203, and the management server 203 drives the directional audio source module 202 to stop pushing advertising audio data N, or drives the directional audio source module 202 to push other suitable advertising audio data N , In order to avoid creating a bad consumer experience for consumers.

請參閱「第9圖」，圖中所示為本發明之另一實施例（二），承「第7圖」所述，為進一步避免重覆推播消費者不喜歡的廣告，本發明的管理伺服器203係具有一生理特徵記錄單元2033，當情緒特徵辨識單元2013辨識出消費者的情緒資訊為負向情緒時，生理特徵記錄單元2033會記錄該名消費者的辨識資訊F以及當下所推播給該名消費者的該則廣告音訊資料N之類型，藉此，有效避免系統下次再度推播同樣類型之廣告音訊資料N給該名消費者，以造成該名消費者的不悅。Please refer to "Figure 9". The figure shows another embodiment (2) of the present invention. Following the description of "Figure 7", in order to further avoid repetitive promotion of advertisements that consumers do not like, the present invention The management server 203 has a physiological characteristic recording unit 2033. When the emotional characteristic recognition unit 2013 recognizes that the consumer’s emotional information is a negative emotion, the physiological characteristic recording unit 2033 will record the consumer’s identification information F and the current situation. The type of the advertisement audio data N pushed to the consumer, thereby effectively preventing the system from pushing the same type of advertisement audio data N to the consumer next time, which will cause the consumer to be unhappy .

由上所述可知，本發明可投放於特定人員的語音廣告系統及其實施方法，其主要係將三維影像感測模組設於室內外環境中，以辨識消費者的生物特徵及屬性，經過管理伺服器的分析及配對，再透過管理伺服器驅動指向性音源模組發送適合該名消費者屬性的廣告音訊資料；依此，本發明其據以實施後，確實可達到提供一種可辨識開放空間範圍內的消費者之生物特徵，藉以推播適合該消費者屬性的廣告訊息，且系統所推播的廣告訊息可在不影響他人的狀況下進行投放的語音廣告系統之目的。From the above, it can be seen that the voice advertisement system and its implementation method of the present invention can be put on specific personnel. It is mainly to install the three-dimensional image sensing module in indoor and outdoor environments to identify the biological characteristics and attributes of consumers. The management server analyzes and matches, and then drives the directional sound source module through the management server to send advertising audio data suitable for the consumer's attributes; according to this, after the present invention is implemented according to it, it can indeed provide a recognizable openness The biological characteristics of consumers in the space range are used to push advertisement messages suitable for the attributes of the consumers, and the advertisement messages promoted by the system can be used for the purpose of a voice advertisement system without affecting others.

唯，以上所述者，僅為本發明之較佳之實施例而已，並非用以限定本發明實施之範圍；任何熟習此技藝者，在不脫離本發明之精神與範圍下所作之均等變化與修飾，皆應涵蓋於本發明之專利範圍內。However, the above are only the preferred embodiments of the present invention, and are not intended to limit the scope of implementation of the present invention; anyone who is familiar with this technique can make equal changes and modifications without departing from the spirit and scope of the present invention , Should be covered in the scope of the patent of the present invention.

綜上所述，本發明之功效，係具有發明之「產業可利用性」、「新穎性」與「進步性」等專利要件；申請人爰依專利法之規定，向　鈞局提起發明專利之申請。To sum up, the effect of the present invention is that it possesses the patent requirements of "industrial applicability", "novelty" and "progressiveness" of the invention; the applicant filed a patent for invention with the Bureau in accordance with the provisions of the Patent Law Application.

10:商場 20:可投放於特定人員的語音廣告系統 201:三維影像感測模組 2011:辨識單元 2012:追蹤單元 2013:情緒特徵辨識單元 202:指向性音源模組 203:管理伺服器 2031:資料庫 2032:族群分類單元 2033:生理特徵記錄單元 S1:影像擷取步驟 S2:消費者屬性比對步驟 S3:廣告音訊配對步驟 S4:投放廣告音訊步驟 A:影像資訊 F:辨識資訊 G:建議推播資訊 H:臉部特徵資訊 N:廣告音訊資料 P:開放空間 10: mall 20: Voice advertising system that can be placed on specific personnel 201: 3D image sensing module 2011: Identification Unit 2012: tracking unit 2013: Emotional Feature Recognition Unit 202: Directional sound source module 203: Management server 2031: database 2032: ethnic taxa 2033: Physiological Characteristic Recording Unit S1: Image capture step S2: Consumer attribute comparison steps S3: Advertising Audio Pairing Steps S4: Steps for advertising audio A: Image information F: Identification information G: Recommended to push information H: Facial feature information N: Advertising audio data P: open space

第1圖，為習知商場的語音廣告推播示意圖。第2圖，為本發明之系統架構圖（一）。第3圖，為本發明之系統架構圖（二）。第4圖，為本發明之實施示意圖（一）。第5圖，為本發明之實施示意圖（二）。第6圖，為本發明之實施示意圖（三）。第7圖，為本發明之實施步驟流程示意圖。第8圖，為本發明之另一實施例（一）。第9圖，為本發明之另一實施例（二）。Figure 1 is a schematic diagram of the voice advertisement push broadcast of Xizhi Shopping Mall. Figure 2 is a system architecture diagram (1) of the present invention. Figure 3 is the system architecture diagram (2) of the present invention. Figure 4 is a schematic diagram (1) of the implementation of the present invention. Figure 5 is a schematic diagram of the implementation of the present invention (2). Figure 6 is a schematic diagram of the implementation of the present invention (3). Figure 7 is a schematic diagram of the process flow of the implementation steps of the present invention. Figure 8 shows another embodiment (1) of the present invention. Figure 9 shows another embodiment (2) of the present invention.

S1:影像擷取步驟 S1: Image capture step

S2:消費者屬性比對步驟 S2: Consumer attribute comparison steps

S3:廣告音訊配對步驟 S3: Advertising Audio Pairing Steps

S4:投放廣告音訊步驟 S4: Steps for advertising audio

Claims

一種可投放於特定人員的語音廣告系統，其包含: 一三維影像感測模組，具有一辨識單元以及一追蹤單元，該三維影像感測模組可對一開放空間內的數名走動的消費者進行影像擷取作業，以產生一影像資訊，該影像資訊可透過該辨識單元進行人體生物特徵的辨識，以產生一辨識資訊，而該追蹤單元可追蹤該影像資訊中移動的人體的位置資訊；一管理伺服器，與該三維影像感測模組形成資訊連線，可接收該辨識資訊，該管理伺服器具有一資料庫及一族群分類單元，該資料庫可儲存數筆廣告音訊資料，而該族群分類單元可分析比對該辨識資訊後產生一建議推播資訊，該管理伺服器可將該建議推播資訊與資料庫中儲存的數筆該廣告音訊資料配對，以挑選出適合該建議推播資訊的該廣告音訊資料；以及一指向性音源模組，與該管理伺服器形成資訊連線，該管理伺服器可搭配該追蹤單元所取得的位置資訊，驅動該指向性音源模組持續對其中一名被標定的消費者進行該廣告音訊資料的投放。A voice advertising system that can be placed on specific people, which includes: A three-dimensional image sensing module has an identification unit and a tracking unit. The three-dimensional image sensing module can perform image capturing operations on several walking consumers in an open space to generate image information. The image information can be used to identify the biological characteristics of the human body through the identification unit to generate identification information, and the tracking unit can track the position information of the moving human body in the image information; A management server forms an information connection with the three-dimensional image sensing module to receive the identification information. The management server has a database and an ethnic classification unit. The database can store several advertisement audio data, and The ethnic classification unit can analyze and compare the identification information to generate a recommended push information, and the management server can match the recommended push information with several pieces of advertising audio data stored in the database to select a suitable recommendation The advertising audio data of the push information; and A directional sound source module forms an information connection with the management server. The management server can cooperate with the location information obtained by the tracking unit to drive the directional sound source module to continuously perform a check on one of the calibrated consumers. The delivery of the advertising audio data.

如申請專利範圍第1項所述之可投放於特定人員的語音廣告系統，其中，該三維影像感測模組主要為一深度攝影機(Depth Camera)。As described in item 1 of the scope of patent application, the voice advertisement system that can be put on specific personnel, wherein the three-dimensional image sensing module is mainly a depth camera.

如申請專利範圍第1項所述之可投放於特定人員的語音廣告系統，其中，該指向性音源模組可為多組同時實施，以同時對多名消費者投放適合其屬性的該廣告音訊資料。As described in item 1 of the scope of patent application, the voice advertisement system that can be placed on specific personnel, wherein the directional sound source module can be implemented in multiple groups at the same time, so that the advertisement information suitable for their attributes can be delivered to multiple consumers at the same time data.

如申請專利範圍第1項所述之可投放於特定人員的語音廣告系統，其中，該管理伺服器的該資料庫中預先儲存有數筆臉部特徵資訊。As described in the first item of the patent application, the voice advertisement system that can be placed on a specific person, wherein the database of the management server is pre-stored with several pieces of facial feature information.

如申請專利範圍第4項所述之可投放於特定人員的語音廣告系統，其中，該三維影像感測模組中設有一情緒特徵辨識單元，透過該情緒特徵辨識單元可辨識出消費者目前的臉部情緒，並與該資料庫中儲存的數筆該臉部特徵資訊進行比對分析，以產生一情緒資訊。As described in item 4 of the scope of patent application, the voice advertisement system that can be placed on specific personnel, wherein the three-dimensional image sensing module is provided with an emotional feature recognition unit, through which the current emotional feature recognition unit can identify the current consumer The facial emotions are compared and analyzed with several pieces of facial feature information stored in the database to generate emotion information.

如申請專利範圍第5項所述之可投放於特定人員的語音廣告系統，其中，該情緒特徵辨識單元主要透過一神經網路(Neural Network)模型訓練出臉部特徵辨識模型，以辨識複數臉部影像中的臉部特徵是否與該資料庫中儲存的數筆該臉部特徵資訊吻合。As described in item 5 of the scope of patent application, a voice advertisement system that can be placed on a specific person, wherein the emotional feature recognition unit mainly trains a facial feature recognition model through a neural network model to recognize plural faces Whether the facial features in the image are consistent with the facial feature information stored in the database.

如申請專利範圍第6項所述之可投放於特定人員的語音廣告系統，其中，該臉部特徵資訊包括性別、年齡以及情緒。As described in item 6 of the scope of patent application, a voice advertisement system that can be placed on a specific person, wherein the facial feature information includes gender, age, and emotion.

如申請專利範圍第7項所述之可投放於特定人員的語音廣告系統，其中，該情緒可包括一正向情緒及一負向情緒其中之一。As described in item 7 of the scope of patent application, the voice advertisement system that can be placed on a specific person, wherein the emotion can include one of a positive emotion and a negative emotion.

如申請專利範圍第8項所述之可投放於特定人員的語音廣告系統，其中，該正向情緒為興奮、快樂、愉悅、輕鬆、寧靜及冷靜其中之一。As described in item 8 of the scope of patent application, the voice advertisement system that can be placed on specific personnel, wherein the positive emotion is one of excitement, happiness, joy, relaxation, tranquility and calmness.

如申請專利範圍第8項所述之可投放於特定人員的語音廣告系統，其中，該負向情緒係為苦惱、生氣、緊張、悲傷、無聊及困倦其中之一。As described in item 8 of the scope of patent application, the voice advertisement system that can be placed on specific personnel, wherein the negative emotion is one of distress, anger, nervousness, sadness, boredom, and drowsiness.

如申請專利範圍第8項所述之可投放於特定人員的語音廣告系統，其中，若該情緒資訊比對結果為該正向情緒，則持續推播該廣告音訊資料，若否，則停止繼續推播該廣告音訊資料。For example, the voice advertisement system that can be placed on specific personnel as described in item 8 of the scope of patent application, wherein if the comparison result of the emotion information is the positive emotion, the advertisement audio data will be continuously pushed; if not, the advertisement audio data will be stopped. Push the advertisement audio data.

如申請專利範圍第1項所述之可投放於特定人員的語音廣告系統，其中，該管理伺服器具有一生理特徵記錄單元。As described in item 1 of the scope of patent application, the voice advertisement system that can be placed on a specific person, wherein the management server has a physiological characteristic recording unit.

一種可投放於特定人員的語音廣告系統的實施方法，其包含: 一影像擷取步驟步驟：利用一三維影像感測模組對一開放空間內的數名消費者進行影像擷取作業，並將所擷取的影像資訊利用一辨識單元進行人體生物特徵的辨識，並在完成辨識後產生多筆辨識資訊；一消費者屬性比對步驟：數筆該辨識資訊分別被傳送至一管理伺服器的一族群分類單元進行比對分類，以產生數筆相對應該辨識資訊的建議推播資訊；一廣告音訊配對步驟：該管理伺服器進一步將數筆該建議推播資訊與一資料庫中儲存的數筆廣告音訊資料進行配對，以挑選出符合該建議推播資訊的該廣告音訊資料；以及一投放廣告音訊步驟：該管理伺服器搭配一追蹤單元所取得的位置資訊，進一步驅動一指向性音源模組持續對其中一名被標定的消費者進行該廣告音訊資料的投放。An implementation method of a voice advertisement system that can be placed on specific personnel, which includes: An image capturing step: using a three-dimensional image sensing module to perform image capturing operations on several consumers in an open space, and using a recognition unit to recognize the biological characteristics of the human body using the captured image information, And generate multiple pieces of identification information after the identification is completed; A consumer attribute comparison step: several pieces of the identification information are respectively sent to a group classification unit of a management server for comparison and classification, so as to generate several pieces of recommended push information corresponding to the identification information; An advertisement audio matching step: the management server further matches several pieces of the recommended push information with several pieces of advertisement audio data stored in a database to select the advertisement audio data that meets the recommended push information; and A step of placing advertising audio: the management server is combined with the location information obtained by a tracking unit to further drive a directional audio source module to continuously deliver the advertising audio data to one of the designated consumers.

如申請專利範圍第13項所述之可投放於特定人員的語音廣告系統的實施方法，其中，該投放廣告音訊步驟中，該三維影像感測模組設有一情緒特徵辨識單元，且該管理伺服器的該資料庫中預先儲存有數筆臉部特徵資訊，當消費者接收推播的該廣告音訊資料後，該三維影像感測模組可擷取該消費者的臉部影像，並進一步透過該情緒特徵辨識單元辨識出該消費者目前的臉部情緒，並與該資料庫中儲存的數筆該臉部特徵資訊進行比對分析，以產生一情緒資訊，若該情緒資訊比對結果為一正向情緒，則持續推播該廣告音訊資料，若該情緒資訊比對結果為一負向情緒，停止對該消費者推播該廣告音訊資料。As described in item 13 of the scope of patent application, the method for implementing a voice advertisement system that can be placed on a specific person, wherein, in the step of placing advertisement audio, the three-dimensional image sensing module is provided with an emotional feature identification unit, and the management server There are several facial feature information pre-stored in the database of the device. After the consumer receives the pushed advertisement audio data, the three-dimensional image sensing module can capture the consumer’s facial image, and further use the The emotion feature recognition unit recognizes the current facial emotion of the consumer, and compares and analyzes the facial feature information stored in the database to generate emotion information. If the emotion information comparison result is a If the sentiment is positive, the advertisement audio data is continuously pushed, and if the result of the comparison of the sentiment information is a negative emotion, the advertisement audio data is stopped to be pushed to the consumer.

如申請專利範圍第14項所述之可投放於特定人員的語音廣告系統的實施方法，其中，該管理伺服器具有一生理特徵記錄單元，當該情緒特徵辨識單元辨識出該消費者的該情緒資訊為該負向情緒時，該生理特徵記錄單元會記錄該名消費者的該辨識資訊以及當下所推播給該名消費者的該則廣告音訊資料之類型，避免下次再度推播同樣類型之該廣告音訊資料給該名消費者。As described in item 14 of the scope of patent application, the method for implementing a voice advertisement system that can be placed on a specific person, wherein the management server has a physiological characteristic recording unit, when the emotional characteristic identification unit recognizes the consumer’s emotion When the information is the negative emotion, the physiological characteristic recording unit will record the identification information of the consumer and the type of the advertisement audio data currently pushed to the consumer, so as to avoid pushing the same type again next time The advertising information is given to the consumer.