A textile fabric classification framework through small motions in videos
- PDF / 534,706 Bytes
- 14 Pages / 439.37 x 666.142 pts Page_size
- 88 Downloads / 217 Views
A textile fabric classification framework through small motions in videos Tao Peng 1 & Xianzi Zhou 1 & Junping Liu 1 & Xinrong Hu 1 & Changnian Chen 1 & Zhonghua Wu 2 & Di Peng 1 Received: 3 December 2019 / Revised: 12 October 2020 / Accepted: 15 October 2020 # Springer Science+Business Media, LLC, part of Springer Nature 2020
Abstract
In the field of computer visions, it is demanding and challenging to determine fabric categories in accordance with appearance changes and multi-frame motion information from a video. Investigating recent impressive results on textile fabric classification techniques, we observed that motion-based video analytics were overlooked in the prior studies. To address this technological gap, a framework called Two-Stream+, which employs deep neural networks to classify textile fabrics through small motions in videos is proposed. At the heart of the Two-Stream+ framework, the motion information of textile (e.g., flow trajectories and dense trajectories) was used to expose the material properties. More specifically, we advocate for fusing spatial and temporal Convolutional Neural Networks (i.e., ConvNets) towers at the first fully connected layer. In addition, deformable convolution is used in Residual Networks (i.e., ResNet) to enhance the transformation modeling capability of ConvNets. Testing a publicly available database, a conducted experiments is used to illustrate that the Two-Stream+ architecture has distinct advantages over the state-of-the-art architectures for classifying textile fabrics. Keywords Textile material properties . Small motion . Two-stream architecture . Dense trajectories
1 Introduction Textile fabrics are unique materials consisting of various material properties. Textile fabric recognition techniques play a significant role in the textile industry. It is a traditional wisdom to conduct official procedure of burn tests at professional workshops to identify unknown * Tao Peng [email protected]
1
School of Mathematics and Computer Science of Wuhan Textile University, Wuhan, China
2
Wuhan Aopu Information Ltd, Wuhan, China
Multimedia Tools and Applications
textile fabrics. The fabric materials are appeased and rightly labeled on the basis of burning fiber smell and melting. The downside of this traditional method is that it inevitably causes man-made damages. Therefore, we advocate for a computerized fabric textile classification method, which increasingly becomes popular in the textile industry as well as in the scientific research community. Additionally, computing-based analysis of fabric classification is very beneficial for a wide range of applications such as robotics, online shopping, 3-D reconstruction, textile material editing, and to name a few. In the field of computer vision, considerable attention has been paid toward fabric textile classification from videos. Previous work in estimating the textile fabric properties are mainly focused on cloth recognition and inference from dynamic scenes [2, 3, 11]. More recently, Yang et al. combined appearance and
Data Loading...
