Method, device and system for intelligent hairstyle recommendation based on neural network and unity
A technology of neural network and recommendation method, applied in the field of intelligent hairstyle recommendation, can solve the problems of single standard, poor experience effect, heavy workload, etc., and achieve high practicability, good prediction effect, and vivid effect
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Method used
Image
Examples
Embodiment 1
[0062] like figure 1 As shown, this embodiment provides an intelligent hairstyle recommendation method based on neural network and Unity, including the following steps:
[0063] S1: Collect the mapping relationship between facial features and hairstyle attributes and establish a facial feature-hairstyle attribute mapping table;
[0064] In this embodiment, the user's facial features are divided into 8 two-category indicators. They are: index 1 - forehead length, index 2 - forehead width, index 3 - atrium length, index 4 - face length, index 5 - face width, index 6 - protruding jawbone, index 7 - eye spacing Too wide, indicator eight - whether it is an oval face. The corresponding relationship between facial features and hairstyles in this embodiment is shown in Table 1 below. There are different data processing methods for different facial features, and different neural network architectures are used for network training.
[0065] Table 1. Mapping relationship between facia...
Embodiment 2
[0107] This embodiment provides an intelligent hairstyle recommendation device based on neural network and Unity, including: a mapping table construction module, a data set construction module, an image preprocessing module, a classification prediction module, a facial feature analysis module, a hairstyle library construction module, and a hairstyle number Output module, user face 3D model building module, 3D hairstyle recommendation effect display model building module and effect display module;
[0108] In the present embodiment, the mapping table building module is used to collect the mapping relationship between facial features and hairstyle attributes and establish a mapping table between facial features and hairstyle attributes, and split the facial features into multiple two-category indexes;
[0109] In this embodiment, the dataset building module is used to construct a dataset for facial feature extraction;
[0110] In this embodiment, the image preprocessing module i...
Embodiment 3
[0120] like image 3 As shown, this embodiment provides an intelligent hairstyle recommendation system based on neural network and Unity, is provided with the intelligent hairstyle recommendation device based on neural network and Unity of The client communication module and the Android client central control module; the server end is also provided with a Socket server communication module, a database module and a server central control module;
[0121] like Figure 4 As shown in the figure, the neural network model is trained based on facial features, and the intelligent hairstyle recommendation algorithm is implemented. In the overall system framework, the facial feature analysis and hairstyle recommendation process are stored in a Python script, and the UI user interface is an interface that provides user interaction with the client. It is used to receive user pictures and additional requirements, such as "occasion", "age", "style"; the Socket client communication module i...
PUM
Login to View More Abstract
Description
Claims
Application Information
Login to View More