Method for recognizing minority clothing images
A technology of minority and recognition methods, applied in character and pattern recognition, instruments, computer parts, etc., to achieve the effect of improving recognition efficiency, clear tone levels, and reducing over-learning problems
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Method used
Image
Examples
Example Embodiment
[0023] Example 1: Such as Figure 1-6 As shown, a method for identifying ethnic minority clothing images, the specific steps are as follows:
[0024] Step1 Perform human body detection on the image G and training image T to be recognized, using k-poselet (k> 1) The deformable component model detects each independent poselet to realize the overall and partial detection of the human body, and obtain the detected image G'and the detected training image T'respectively;
[0025] Step2 Extract the five underlying features of the color histogram, HOG, LBP, SIFT, and edge operator of the detected image G'and the detected training image T'to be recognized, respectively, to obtain the to-be-recognized image G" and Training image T after feature extraction";
[0026] Step3: Define the semantic attributes of ethnic minority clothing, label the detected training images T'with semantic attributes, use multi-task feature models to learn different styles of ethnic minority clothing and train classif...
Example Embodiment
[0031] Embodiment 2: In this embodiment, an image of Yunnan ethnic minority clothing is taken as an example for description.
[0032] Step1. First input the ethnic clothing image G to be recognized and the training image T from the Yunnan ethnic clothing image database, using the weight vector ω=(M 0 ,...,M j ...,M k-1 ,d 1 ,...,d j ...d k-1 ,b) Describe each k-poselet, where M j Is the appearance template, d j It is the spatial deformation model of the j-th pose of k-poselet, and b is the deviation. When the model is detected, each k-poselet is described by a weight vector.
[0033] Then, k individual HOG templates are used to simulate the appearance model of each part, the human body detection is performed on each independent poselet, and the key point prediction is made from the poselet position and the average position of the scale in the training data. Use average maximum accuracy (AMP) to measure whether a set of k-poselets C achieves high accuracy and high coverage:
[0034] ...
PUM
Abstract
Description
Claims
Application Information
- R&D Engineer
- R&D Manager
- IP Professional
- Industry Leading Data Capabilities
- Powerful AI technology
- Patent DNA Extraction
Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.
© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap