The automatic recognition of garment flat information has been widely researched through computer vision. However, the unapparent visual feature and low recognition accuracy pose serious challenges to the application. Herein, inspired by multi-object instance segmentation, the method of mask region convolutional neural network (Mask R-CNN) for garment flat multi-component is proposed in this paper. The steps include feature enhancement, attribute annotation, feature extraction, and bounding box regression and recognition. First, the Laplacian was employed to enhance the image feature, and the Polygon annotated component attributes to reduce the interaction interference. Next, the ResNet was applied to realize identity mapping to characterize redundant information of components. Finally, the feature map was entered into two branches to achieve bounding box regression and recognition. The results demonstrated that the proposed method could realize multi-component recognition effectively. Compared with the unenhanced feature, the mAP increased by 2.27%, reaching 97.87%, and the average F1 was 0.958. Compared to VGGNet and MobileNet, the ResNet backbone used for Mask R-CNN could improve the mAP by 11.55%. Mask R-CNN was more robust than the state-of-the-art methods and more suitable for garment flat multi-component recognition.

Clothing matching refers to the coordination of style, color, etc. to achieve a decent and generous effect. With the development of artificial intelligence, increasing research efforts have been dedicated to complementary garment collocation as matching clothes to make a suitable outfit has become a daily headache for many people. However, existing studies neglect rules regarding clothing matching, which are based on the knowledge accumulated in the fashion domain and are implicitly transmitted and ambiguous. Towards this end, this article proposes an expert system based on domain knowledge implemented using the Prolog program to realize rule-guided clothing collocation. The article constructs clothing matching rules from four essential clothing attributes: season, type, style, and color, and uses Prolog syntax for knowledge representation. For the formulated facts and rules, the reasoning machine iteratively matches the user's instructions with the system knowledge to reason out suitable matching suggestions. In this study we built a website as a man-machine interface to facilitate friendly interaction and set season, type, style, and color sub-options to match clothing and satisfy user preferences. The system validation based on standard metrics (precision, recall, F1-measure) achieves results above 81%. The main contribution is that the system could match clothes more accurately according to the likes and requirements of users. After comparing with other research, the system is expected to provide suggestions or references for people to choose style and dressing.

PurposeConsidering only two-dimensional (2D) ease allowance cannot fully reflect the three-dimensional (3D) relationship between the position of clothing and the human body. The purpose of this paper is to propose a method with a 3D space vector and corresponding distance ease to characterize fitting garments and then used to construct personalized clothing for similar shape body.Design/methodology/approachFirstly, a 3D scanner was used to obtain mannequin and fitted garment data, and 17 layers of cross-sections of the upper body were extracted. Then, 37 space vectors and corresponding space angles on each cross-section were obtained with the original point. Secondly, the detailed distance ease between the mannequin and garment was constructed due to the difference between garment vectors and body vectors. Thirdly, the distance ease mathematical models were achieved and used to calculate distance ease on a similar shape body. Additionally, the fit garment is constructed, and the garment pattern is altered by the geometric pattern alteration method.FindingsThe results show that 3D space vectors can explain the relationship between body skin and garment surface of the upper body properly. The distance ease is modeled by mathematic expressions and successfully used to make a new garment to fit a similar shape body.Originality/valueThe proposed method of constructing garments based on distance ease and 3D space vectors can create a fitted garment for a similar shape body effectively and accurately. It is useful for the personalized garment design and suitable for the manufacturing process.