TWI838887B - Artificial intelligence importing personal photo background color recommendation system and method thereof - Google Patents

Abstract

The invention discloses an artificial intelligence importing personal photo background color recommendation system and method thereof. A color capture module captures a user's body part color and/or wearing object color from a personal photo image file with the user's personal photo, and then the color capture module inputs the user's body part color and/or wearing object color into a deep learning model of an artificial intelligence module. Next, the deep learning model of the artificial intelligence module generates a recommended background color according to the user's body part color and/or wearing object color input by the color capture module, and then an image processing module replaces or updates a background color of the personal photo of the personal photo image file with the recommended background color by using an image processing method, so as to deepen or reinforce a viewer's favorable impression for the user or the personal photo.

Description

Artificial intelligence introduced into personal photo background color recommendation system and method

The present invention relates to a personal photo background color recommendation technology, and in particular to an artificial intelligence (AI)-introduced personal photo background color recommendation system and method thereof.

The user's personal photos (such as headshots/upper body photos/full body photos) can be widely used as resume (job search) photos, social photos, further education (enrollment) photos, network photos, election photos, business card photos, etc., and can also be used in various purposes such as job search resumes, human resources banks, dating software, promotional (advertising) websites, election billboards, personal services, etc., and research reports point out that viewers only need to look at unfamiliar faces in personal photos for a short time (such as 100 milliseconds) to form an impression and feature judgment, so personal photos are very important for different viewers such as interviewers, dating targets, Internet users, voters, etc. to leave a good impression.

Taking the example of using personal photos as resume photos, according to the description of job search websites (such as 104 or 1111 job banks), company HR managers often have to screen out potential job seekers from a large number of resumes, and the personal photo (resume photo) in the resume can easily become an important factor for HR managers to decide whether to interview or consider job seekers.

In addition, most personal photos today are taken by photo studios or by users themselves, but there is currently no relevant research on how to optimize the background color of personal photos to enhance the viewer's visual impression, nor is there any relevant technology to improve or recommend methods for the production of background colors for personal photos.

Furthermore, the different impressions or images that personal photos give to viewers can be traced back to the theory of using color systems or color science. Color systems or color science define colors in a scientific and quantitative way to facilitate users to apply them accurately and objectively. However, various color systems have different industry characteristics or definition methods and have developed different advantages and characteristics.

In addition, color harmony refers to the matching of two or more colors. The phenomenon of color harmony can produce order and harmony to make people feel pleasant. The feeling of color is related to each person's subjective consciousness, different cultures, external environment or life experience background, but for people who have not received professional color training, it is relatively complicated and difficult to use the theory and method of color harmony. Therefore, there are many color matching analysis tools available on the Internet, but these color matching analysis tools are very difficult for non-professionals to operate, and it is impossible to confirm whether they can improve the effect of color matching or color harmony.

Therefore, how to provide an innovative personal photo background color recommendation technology to solve any of the above problems or provide related functions/services has become a major research topic for technical personnel in this field.

The present invention provides an innovative artificial intelligence-introduced personal photo background color recommendation system and method, which can input the body part color and/or clothing color of the user in the personal photo into the deep learning model of the artificial intelligence module to generate a recommended background color, and can also use the recommended background color to optimize the background color of the user's personal photo to improve the viewer's satisfaction with the user, or generate different recommended background colors through the deep learning model of the artificial intelligence module or the neural network of multi-layer sensors to replace, update or optimize the background colors of different personal photos, or the color combination collection module automatically collects multiple color combination data sets through the color combination database and/or web crawler technology, or the data training module trains the deep learning model of the artificial intelligence module or the neural network of multi-layer sensors.

The artificial intelligence-introduced personal photo background color recommendation system of the present invention comprises: a color capture module, which captures at least one of the body part color and the color of the clothing of the user from a personal photo image file having the personal photo of the user; an artificial intelligence module with a deep learning model, which is interconnected or communicated with the color capture module, so that the color capture module inputs at least one of the body part color and the color of the clothing of the user in the personal photo into the deep learning model of the artificial intelligence module, and then the deep learning model of the artificial intelligence module is used to recommend the background color of ... A recommended background color is generated based on at least one of the body part color and the color of the clothing of the user in the personal photo input by the color capture module; and an image processing module is interconnected or communicated with the artificial intelligence module having a deep learning model, so that the image processing module replaces or updates the background color of the personal photo in the personal photo image file with the recommended background color generated by the deep learning model of the artificial intelligence module based on at least one of the body part color and the color of the clothing of the user in the personal photo input by the color capture module through image processing.

The method for recommending background color of a personal photo by artificial intelligence includes: extracting at least one of the color of a user's body part and the color of an item of clothing from a personal photo image file having the user's personal photo by a color capture module, inputting at least one of the color of the user's body part and the color of an item of clothing in the personal photo into a deep learning model of an artificial intelligence module by the color capture module; and A recommended background color is generated based on at least one of the body part color and the color of the clothing of the user in the personal photo input by the color capture module, and the image processing module replaces or updates the background color of the personal photo in the personal photo image file with the recommended background color generated by the deep learning model of the artificial intelligence module based on at least one of the body part color and the color of the clothing of the user in the personal photo input by the color capture module through image processing.

In order to make the above features and advantages of the present invention more clearly understandable, the following examples are given and detailed descriptions are provided in conjunction with the attached drawings. The following description will partially explain the additional features and advantages of the present invention, and these features and advantages will be partially known from the description or can be learned through the practice of the present invention. It should be understood that both the general description above and the detailed description below are exemplary and explanatory, and are not intended to limit the scope of the present invention.

1: Artificial intelligence is introduced into the personal photo background color recommendation system

2: Background color recommendation subsystem

11: Receiving module

12: Color capture module

13: Artificial intelligence module

14: Image processing module

15: Background color database

16: Display module

21: Color combination database

22: Color combination collection module

23: Color data quantization module

24: Data training module

25: Application Programming Interface

A: Personal photo image file

A1: Personal photo

A2: Background color

B1: Body part color

B2: Color of clothing

C: Deep learning model

C': Neural network with multiple layers of sensors

C1: Input layer

C2: Hidden layer

C3: Output layer

D1: Recommended background color

D2: Candidate background color

E: Image processing method

F: Web crawler technology

G: Color combination data set

H: Training data

S1 to S2: Steps

Figure 1 is a schematic diagram of the architecture of the artificial intelligence-introduced personal photo background color recommendation system of the present invention.

Figure 2 is another schematic diagram of the artificial intelligence-introduced personal photo background color recommendation system of the present invention.

Figure 3 is a schematic diagram of a process of introducing artificial intelligence into a personal photo background color recommendation method of the present invention.

FIG4 is a schematic diagram of a deep learning model (e.g., a neural network of multi-layer perceptrons) in the artificial intelligence-introduced personal photo background color recommendation system and method of the present invention.

The following describes the implementation of the present invention through a specific concrete implementation form. People familiar with this technology can understand other advantages and effects of the present invention from the content disclosed in this manual, and can also implement or use it through other different specific equivalent implementation forms.

FIG1 is a schematic diagram of the structure of the artificial intelligence-introduced personal photo background color recommendation system 1 of the present invention, FIG2 is another schematic diagram of the structure of the artificial intelligence-introduced personal photo background color recommendation system 1 of the present invention, and FIG3 is a schematic diagram of the process of the artificial intelligence-introduced personal photo background color recommendation method of the present invention.

As shown in FIG1 , the artificial intelligence-introduced personal photo background color recommendation system 1 mainly includes a receiving module 11, a color capture module 12, an artificial intelligence module 13, an image processing module 14, a background color database 15 and a display module 16 that are connected or communicated with each other, and the receiving module 11, the color capture module 12, the artificial intelligence module 13, the image processing module 14, the background color database 15 and the display module 16 can first form a background color recommendation subsystem 2. For example, the receiving module 11 can be connected or communicated with the color capture module 12, the image processing module 14 and the display module 16, respectively, and the artificial intelligence module 13 can be connected or communicated with the color capture module 12, the image processing module 14 and the background color database 15, respectively.

As shown in FIG2 , the artificial intelligence-introduced personal photo background color recommendation system 1 may further include a color combination database 21, a color combination collection module 22, a color data quantization module 23, a data training module 24, and an application programming interface (API) 25, etc., which are interconnected or communicated. For example, the color combination collection module 22 may be connected or communicated with the color combination database 21 and the color data quantization module 23, respectively, the data training module 24 may be connected or communicated with the color data quantization module 23 and the background color recommendation subsystem 2 (see FIG1 ), respectively, and the background color recommendation subsystem 2 may be further connected or communicated with the application programming interface (API) 25.

In one embodiment, the receiving module 11 may be a receiver (chip/circuit), receiving software (program), etc., and the color capture module 12 may be a color capturer (chip/circuit), color capture software (program), etc. The artificial intelligence module 13 may be an artificial intelligence chip (circuit), artificial intelligence software (program), etc., and the deep learning model C of the artificial intelligence module 13 may be a neural network C' of a multilayer perceptron (MLP) (see FIG. 4 ). The image processing module 14 may be an image processor (chip/circuit), image processing software (program), etc. The background color database 15 may be a server, storage, memory, memory card, hard disk, etc. for storing data of multiple background colors A2, recommended background colors D1 and/or candidate background colors D2. The display module 16 may be a display, display panel, display software (program), etc.

In one embodiment, the color combination database 21 may be a server, storage, memory, memory card, hard disk, etc. for storing data of a plurality of color combination data sets G (such as color matching data sets), the color combination collection module 22 may be a color combination collector (chip/circuit), color combination collection software (program), etc., the color data quantization module 23 may be a color data quantizer (chip/circuit), color data quantization software (program), etc., the data training module 24 may be a data trainer (chip/circuit), data training software (program), etc., and the application programming interface (API) 25 may be an application programming interface, an application development interface, etc.

In one embodiment, the personal photo A1 (or personal photo) can be a headshot (head photo), upper body photo or full body photo of a single user, and the personal photo A1 can be widely used as a resume (job search) photo, a dating photo, a study photo, an online photo, an election photo, a business card photo, etc. It can also be used for various purposes such as job search resumes, job banks, dating software, promotional (advertising) websites, election billboards, personal services, etc. The background color A2 of the personal photo A1 can be a single background color or multiple background colors of the personal photo A1, and the multiple background colors can be a combination of at least two (such as two, three or more than four) background colors.

In one embodiment, the body part color B1 may be the color of the hair part (such as the hair/hair dye color), the face part color (such as the eyes/eye shadow/eyelashes/blush/lip color), the skin part color (such as the skin/makeup/tattoo color of the chest/back/hands/feet), etc. of the user in the personal photo A1. The clothing color B2 may be the color of the clothing (such as clothes/pants/shoes) or the color of the accessories (such as hats/bags/necklaces/watches) worn by the user in the personal photo A1. The background color A2 of the personal photo A1, the body part color B1, the clothing color B2, the recommended background color D1 or the candidate background color D2 may be, for example, a single color or a combination of multiple colors such as red, orange, yellow, green, blue, indigo, purple, gold, silver, copper, black, white, gray, etc., that is, a combination of one color or two or more colors.

In one embodiment, the "connection or communication" described in the present invention may represent interconnection or communication in a wired manner (such as a wired network) or a wireless manner (such as a wireless network), "at least one" represents more than one (such as one, two or more than three), and "plurality" represents more than two (such as two, three, four, five or more than ten). However, the present invention is not limited to those mentioned in the above embodiments.

The artificial intelligence-introduced personal photo background color recommendation system 1 and its method of the present invention may include the contents shown in steps S1 to S2 of Figures 1, 2 and 3.

In step S1, the color capture module 12 captures at least one of the user's body part color B1 and the wear color B2 from the personal photo image file A having the user's personal photo A1, so that the color capture module 12 inputs at least one of the user's body part color B1 and the wear color B2 in the personal photo A1 into the deep learning model C of the artificial intelligence module 13.

In step S2, the deep learning model C of the artificial intelligence module 13 generates a recommended background color D1 based on at least one of the user's body part color B1 and the wear color B2 in the personal photo A1 input by the color capture module 12, and the image processing module 14 replaces or updates the background color A2 of the personal photo A1 of the personal photo image file A with the recommended background color D1 generated by the deep learning model C of the artificial intelligence module 13 based on at least one of the user's body part color B1 and the wear color B2 in the personal photo A1 input by the color capture module 12 through the image processing method E.

In other words, the receiving module 11 may first receive a personal photo image file A having a personal photo A1 of the user, and the color capture module 12 may capture at least one of the user's body part color B1 (such as the color of the hair/face/skin part) and the color of the clothing B2 (such as the color of clothing/accessories) from the personal photo image file A having the personal photo A1 received by the receiving module 11, and then the color capture module 12 may input at least one of the user's body part color B1 and the color of the clothing B2 into the deep learning model C of the artificial intelligence module 13, so that the deep learning model C of the artificial intelligence module 13 may generate (output) a recommended background color D1 according to at least one of the user's body part color B1 and the color of the clothing B2 input by the color capture module 12.

Next, the image processing module 14 replaces (substitutes) or updates the background color A2 of the personal photo A1 of the personal photo image file A with the recommended background color D1 generated (output) by the deep learning model C of the artificial intelligence module 13 through the image processing method E (such as image processing technology/image processing algorithm), and then the display module 16 displays the recommended background color D1 generated by the deep learning model C of the artificial intelligence module 13 by the image processing module 14. The personal photo A1 of the personal photo image file A formed after the background color D1 replaces (replaces) or updates the background color A2 is used by the user to apply the personal photo image file A with the personal photo A1 formed after the background color D1 replaces (replaces) or updates the background color A2 to various purposes such as job application resumes, job banks, dating software, further education (enrollment) websites, promotional (advertising) websites, election billboards, personal services, etc.

In one embodiment, a user can input a personal photo image file A having a personal photo A1 into an artificial intelligence into a personal photo background color recommendation system 1 through an electronic device (such as a smart phone/computer/server, etc.), and then the deep learning model C of the artificial intelligence module 13 generates a recommended background color D1 according to at least one of the user's body part color B1 and the wearable color B2 in the personal photo A1, so that the image processing module 14 sets the recommended background color D1 to the background color A2 of the personal photo A1 through an image processing method E (such as image processing technology/image processing algorithm), thereby automatically recommending the recommended background color D1 that best matches the personal photo A1 to optimize the background color A2 of the personal photo A1.

In one embodiment, the artificial intelligence module 13 may generate a candidate background color D2 in advance based on the user's body part color B1 and the clothing color B2 in the personal photo A1, so that the artificial intelligence module 13 updates the candidate background color D2 to the deep learning model C to replace it as the recommended background color D1, so that the image processing module 14 can use the image processing method E to replace the candidate background color D2 to set the recommended background color D1 on the background color A2 of the personal photo A1.

In one embodiment, the artificial intelligence module 13 may activate or execute a color suggestion feedback method to input common or frequently used recommended background colors D1 into the deep learning model C, and then the artificial intelligence module 13 uses the common or frequently used recommended background colors D1 to dynamically update the deep learning model C or its background color training sub-model.

As shown in FIG. 1 and FIG. 2, the deep learning model C of the artificial intelligence module 13 can automatically generate a recommended background color D1 based on at least one of the body part color B1 and the clothing color B2 of the user in the personal photo A1. The color combination collection module 22 can collect the recommended background color D1 through the color combination database 21 and the web crawler (web At least one of the crawler technologies F automatically collects multiple color combination data sets G (such as color matching data sets), and the data training module 24 can train the multi-layer perceptron neural network C' used by the deep learning model C of the artificial intelligence module 13, so that the multi-layer perceptron neural network C' used by the deep learning model C of the artificial intelligence module 13 generates different recommended background colors D1 to optimize the background colors A2 of different personal photos A1, which is conducive to deepening or strengthening the viewer's impression of the user or his personal photo A1 through the recommended background color D1.

After the data training module 24 trains the multi-layer sensor neural network C' used by the deep learning model C of the artificial intelligence module 13, the multi-layer sensor neural network C' can generate a recommended background color D1 based on at least one of the user's body part color B1 and the clothing color B2 in the personal photo A1, so that the image processing module 14 can apply the recommended background color D1 generated by the multi-layer sensor neural network C' to the background color A2 of the personal photo A1 through the image processing method E.

FIG4 is a schematic diagram of the artificial intelligence-introduced personal photo background color recommendation system 1 and its method related to the deep learning model C (e.g., a neural network C' of multi-layer perceptrons) of the present invention, and is explained with reference to FIG1 to FIG2.

As shown in FIG4 , the deep learning model C of the artificial intelligence module 13 can adopt the architecture of a multi-layer perceptron neural network C' to train the collected multiple color combination data set G (such as a color matching data set or a big data color application data), and the multi-layer perceptron neural network C' can sequentially include at least one input layer (Input Layer) C1, multiple hidden layers (Hidden Layer) C2 and at least one output layer (Output Layer) C3. For example, the weighting operation of the multi-layer perceptron neural network C' may use an input layer C1 including multiple (e.g., 12) neural cells and an output layer C3 including multiple (e.g., 3) neural cells, and the regression prediction learning between the input layer C1 and the output layer C3 may use multiple (e.g., 4) hidden layers C2, and each hidden layer C2 may include multiple (e.g., 1024) neural cells, but is not limited to this.

As shown in FIG2 , the color combination collection module 22 can automatically collect the color swatch data (such as 44986 pieces of five-color swatch data) of the plurality of color combination data sets G (such as color matching data sets) through at least one of the color combination database 21 and the web crawler technology F, and then the color data quantization module 23 quantifies the color swatch data (such as 44986 pieces of five-color swatch data) of the plurality of color combination data sets G (such as color matching data sets) collected by the color combination collection module 22 to provide a predetermined number or type of color swatches (such as 130 color swatches/colors).

The data training module 24 can use the machine learning algorithm to train the color swatch data (such as 44986 pieces of five-color swatch data) of the multiple color combination data sets G (such as the color matching data set) collected by the color combination collection module 22 to generate multiple training data (such as a total of 112925 pieces of training data H), and then construct the required deep learning model C through the neural network C' of the multi-layer sensor. The characteristic data of various color combination data sets G (such as the color matching data set) in practical applications are difficult to describe easily, and the color coordination of the color combination data set G (such as the color matching data set) has a very high variability and is mixed with various complex factors.

As shown in Figures 1, 2 and 4, the artificial intelligence module 13 can use the architecture of the neural network C' of the multi-layer sensor to train the deep learning model C based on the color swatch data (such as 44986 five-color swatch data) of the complex color combination data set G (such as the color matching data set), so that when the artificial intelligence module 13 obtains the characteristics of the complex colors (color signals), it can predict new color combinations (such as new optimal color combinations) through the deep learning model C. When the artificial intelligence module 13 obtains a new color combination, the trained deep learning model C can use the neural network C' of the multi-layer sensor to predict the best or most suitable recommended background color D1 according to the new color combination.

For example, the artificial intelligence module 13 may input two known colors into the neural network C' of the multi-layer sensor through the deep learning model C, so that the neural network C' of the multi-layer sensor recommends a new third color based on the known two colors to predict the recommended background color D1 of the new three-color combination. Alternatively, the artificial intelligence module 13 may input three known colors into the neural network C' of the multi-layer sensor through the deep learning model C, so that the neural network C' of the multi-layer sensor recommends a new fourth color based on the known three colors to predict the recommended background color D1 of the new four-color combination. Alternatively, the artificial intelligence module 13 can input the known four colors into the multi-layer perceptron neural network C' through the deep learning model C, so that the multi-layer perceptron neural network C' recommends a new fifth color based on the known four colors to predict the recommended background color D1 of the new five-color combination.

The system interface of the artificial intelligence imported personal photo background color recommendation system 1 can provide an application program interface (API) 25, so that the user's electronic device can transmit or upload the personal photo image file A with the personal photo A1 to the receiving module 11 of the artificial intelligence imported personal photo background color recommendation system 1 through the application program interface (API) 25, and the color capture module 12 can capture at least one of the user's body part color B1 and the wearable color B2 from the personal photo image file A with the personal photo A1 received by the receiving module 11, and then the color capture module 12 can capture at least one of the user's body part color B1 and the wearable color B2 from the personal photo image file A with the personal photo A1 received by the receiving module 11. The module 12 inputs at least one of the user's body part color B1 and the wearable color B2 into the deep learning model C of the artificial intelligence module 13, and then the deep learning model C of the artificial intelligence module 13 generates a recommended background color D1 according to at least one of the user's body part color B1 and the wearable color B2, so that the image processing module 14 replaces (replaces) or updates the background color A2 of the personal photo A1 of the personal photo image file A with the recommended background color D1 generated by the deep learning model C of the artificial intelligence module 13 through the image processing method E.

In summary, the artificial intelligence-introduced personal photo background color recommendation system and method of the present invention has at least the following features, advantages or technical effects.

1. The color capture module of the present invention can automatically input the body part color and/or the color of the clothing of the user in the personal photo into the deep learning model of the artificial intelligence module, so that the deep learning model can automatically generate a recommended background color based on the body part color and/or the color of the clothing of the user.

2. The present invention can optimize the background color of the user's personal photo using the recommended background color, so as to improve the viewer's satisfaction with the user or his personal photo.

3. The deep learning model or multi-layer sensor neural network of the artificial intelligence module of the present invention can generate different recommended background colors to replace, update or optimize the background colors of different personal photos, which is beneficial to deepen or strengthen the viewer's good impression of the user or his personal photo through the recommended background color.

4. The image processing module of the present invention can provide an image processing method (such as image processing technology/image processing algorithm) to automatically replace or update the background color of the personal photo in the personal photo image file with the recommended background color generated by the deep learning model of the artificial intelligence module.

5. The artificial intelligence module of the present invention can generate candidate background colors based on the colors of the user's body parts and the colors of the items he/she wears in the personal photo, and automatically update the candidate background colors to the deep learning model to replace them as recommended background colors. It can also automatically apply the recommended background colors replaced by the candidate background colors to the background colors of the personal photo.

6. The present invention can activate or execute the color suggestion feedback method by the artificial intelligence module to facilitate the input of common or frequently used recommended background colors into the deep learning model. The artificial intelligence module can also dynamically update the deep learning model or its background color training sub-model using common or frequently used recommended background colors.

7. The deep learning model of the artificial intelligence module of the present invention can use a neural network with multiple layers of sensors having input layers, hidden layers, and output layers to automatically train a plurality of color combination data sets (such as color matching data sets) using the neural network with multiple layers of sensors.

8. The color combination collection module of the present invention can automatically collect multiple color combination data sets (such as color matching data sets) through the color combination database and/or web crawler technology, and the data training module can automatically train the deep learning model of the artificial intelligence module or the neural network of the multi-layer sensor to generate different recommended background colors to optimize the background colors of different personal photos.

9. The present invention can be used as a personal photo optimization system, optimization method or optimization tool, and can also automatically generate recommended background colors for personal photos based on the characteristics of different users' body parts colors and/or clothing colors, so as to enhance viewers' good impression of the user or his or her personal photos (such as credibility, affinity, professional ability).

10. This invention can be applied to industries or products such as job banks, job search websites, dating software (dating websites), personal advertising, and personal photo optimization applications. It can also be used in many fields such as job hunting, dating, further education (enrollment), elections, and advertising, and has great industrial applicability and commercial practical value.

The above implementation forms are only illustrative of the principles, features and effects of the present invention, and are not intended to limit the scope of implementation of the present invention. Anyone familiar with this technology can modify and change the above implementation forms without violating the spirit and scope of the present invention. Any equivalent changes and modifications completed using the content disclosed by the present invention should still be covered by the scope of the patent application. Therefore, the scope of protection of the present invention should be as listed in the scope of the patent application.

1: Artificial intelligence is introduced into the personal photo background color recommendation system

2: Background color recommendation subsystem

11: Receiving module

12: Color capture module

13: Artificial intelligence module

14: Image processing module

15: Background color database

16: Display module

A: Personal photo image file

A1: Personal photo

A2: Background color

B1: Body part color

B2: Color of clothing

C: Deep learning model

D1: Recommended background color

D2: Candidate background color

E: Image processing method

Claims (10)

An artificial intelligence-introduced personal photo background color recommendation system includes: a color capture module that captures at least one of the body part color and the clothing color of the user from a personal photo image file having the user's personal photo; an artificial intelligence module having a deep learning model composed of a neural network of multiple layers of sensors sequentially including at least one input layer, a plurality of hidden layers and at least one output layer, which is interconnected or communicated with the color capture module to extract at least one of the body part color and the clothing color of the user from the personal photo image file; The color capture module inputs at least one of the body part color and the clothing color of the user in the personal photo of the personal photo image file into the deep learning model of the artificial intelligence module composed of the neural network of the multi-layer sensor including the at least one input layer, the plurality of hidden layers and the at least one output layer in sequence, and then the deep learning model of the artificial intelligence module composed of the neural network of the multi-layer sensor including the at least one input layer, the plurality of hidden layers and the at least one output layer in sequence is used. The learning model generates a recommended background color for replacing or updating the background color of the personal photo of the personal photo image file according to at least one of the body part color and the clothing color of the user in the personal photo input by the color capture module; and an image processing module, which is interconnected with the artificial intelligence module of the deep learning model composed of the neural network of the multi-layer sensor including the at least one input layer, a plurality of hidden layers and at least one output layer in sequence. or communication, so that the image processing module replaces or updates the background color of the personal photo in the personal photo image file with the recommended background color generated by the deep learning model of the artificial intelligence module, which is composed of the neural network of the multi-layer sensor including the at least one input layer, a plurality of hidden layers and the at least one output layer in sequence, according to at least one of the body part color and the clothing color of the user in the personal photo input by the color capture module through image processing. As described in claim 1, the artificial intelligence module generates a candidate background color based on the color of the user's body parts and the color of the clothing in the personal photo, and the artificial intelligence module updates the candidate background color to the deep learning model to replace it as the recommended background color, and then the image processing module applies the recommended background color replaced by the candidate background color to the background color of the personal photo through the image processing method. As described in claim 1, the artificial intelligence module further activates or executes a color suggestion feedback method to input common or frequently used recommended background colors into the deep learning model, and then the artificial intelligence module dynamically updates the deep learning model or its background color training sub-model using the common or frequently used recommended background colors. As described in claim 1, the artificial intelligence is introduced into the personal photo background color recommendation system, wherein the deep learning model of the artificial intelligence module further adopts the neural network of the multi-layer sensor to train the collected multiple color combination data set. As described in claim 1, the artificial intelligence is introduced into the personal photo background color recommendation system, wherein the artificial intelligence module further uses the neural network of the multi-layer sensor to train the deep learning model based on the color swatch data of the multiple color combination data set, so that the artificial intelligence module can predict new color combinations through the deep learning model when obtaining the characteristics of multiple colors, and then the deep learning model uses the neural network of the multi-layer sensor to predict the best or most suitable recommended background color based on the new color combination. The artificial intelligence-introduced personal photo background color recommendation system as described in claim 1 further includes a receiving module and a display module, wherein the receiving module receives a personal photo image file having the personal photo of the user, and the color capture module extracts at least one of the body part color and the clothing color of the user from the personal photo image file having the personal photo received by the receiving module, and the display module displays the personal photo of the personal photo image file formed after the image processing module replaces or updates the background color with the recommended background color generated by the deep learning model of the artificial intelligence module. The artificial intelligence-introduced personal photo background color recommendation system as described in claim 1 further includes a color combination collection module and a data training module, wherein the color combination collection module automatically collects multiple color combination data sets through at least one of a color combination database and a web crawler technology, and the data training module trains the multi-layer perceptron neural network adopted by the deep learning model of the artificial intelligence module, so that the multi-layer perceptron neural network adopted by the deep learning model of the artificial intelligence module generates different recommended background colors to optimize the background colors of different personal photos. As described in claim 7, the artificial intelligence-introduced personal photo background color recommendation system, wherein the data training module trains the multi-layer sensor neural network used by the deep learning model of the artificial intelligence module, and then the multi-layer sensor neural network generates the recommended background color based on at least one of the body part color and the clothing color of the user in the personal photo, and the image processing module applies the recommended background color generated by the multi-layer sensor neural network to the background color of the personal photo through the image processing method. The artificial intelligence-introduced personal photo background color recommendation system as described in claim 1 further includes a color combination collection module and a color data quantization module, wherein the color combination collection module automatically collects color swatch data of a plurality of color combination data sets through at least one of a color combination database and a web crawler technology, and then the color data quantization module quantifies the color swatch data of the plurality of color combination data sets collected by the color combination collection module to provide a predetermined number or type of color swatches. A method for recommending background color of a personal photo by introducing artificial intelligence includes: extracting at least one of the color of a user's body part and the color of an item of clothing from a personal photo image file having the user's personal photo by a color capture module, inputting at least one of the color of the user's body part and the color of an item of clothing in the personal photo of the personal photo image file into a deep learning model of an artificial intelligence module composed of a neural network of a multi-layer sensor including at least one input layer, a plurality of hidden layers and at least one output layer in sequence by the color capture module; and The deep learning model generates a recommended background color for replacing or updating the background color of the personal photo of the personal photo image file according to at least one of the body part color and the color of the clothing of the user in the personal photo input by the color capture module, and the deep learning model of the artificial intelligence module composed of the neural network of the multi-layer sensor including the at least one input layer, a plurality of hidden layers and at least one output layer in sequence generates the recommended background color according to at least one of the body part color and the color of the clothing of the user in the personal photo input by the color capture module to replace or update the background color of the personal photo of the personal photo image file.

Images