Open AccessLandmark‐free clothes recognition with a two‐branch feature selective network
Author(s) -
Lee S.,
Eun H.,
Oh S.,
Kim W.,
Jung C.,
Kim C.
Publication year - 2019
Publication title -
electronics letters
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.375
H-Index - 146
eISSN - 1350-911X
pISSN - 0013-5194
DOI - 10.1049/el.2019.0660
Subject(s) - landmark , discriminative model , computer science , benchmark (surveying) , artificial intelligence , pattern recognition (psychology) , feature (linguistics) , representation (politics) , clothing , set (abstract data type) , task (project management) , computer vision , engineering , cartography , geography , linguistics , philosophy , archaeology , systems engineering , politics , political science , law , programming language
In this Letter, the authors present a ‘landmark‐free’ clothes recognition approach. Recent studies have shown that the use of landmark information has achieved great success in the task of clothes recognition. However, the landmark annotation is very labour intensive and time consuming. It also suffers from inter‐ and intra‐individual variability. To overcome these problems, the authors propose a two‐branch feature selective network for category classification and attribute prediction. Note that, in this Letter, they prove that the proposed network has an excellent ability to effectively learn a discriminative feature representation of a ‘clothing image’. Experimental results on the benchmark data set show that the proposed network yields comparable performance to the state‐of‐the‐art methods, which strongly depend on the fashion landmark.