Sketched symbol recognition using Latent-Dynamic Conditional Random Fields and distance-based clustering
作者：
Highlights：
• A two stage methodology for the recognition of sketched symbols is proposed.
• The approach combines a discriminative graphical model and a clustering algorithm.
• The LDCRF model is applied to the domain of sketched symbol recognition.
• A complete evaluation of the approach in the electric circuit domain is provided.
• The achieved performance is compared with that of a baseline system.
摘要
Highlights•A two stage methodology for the recognition of sketched symbols is proposed.•The approach combines a discriminative graphical model and a clustering algorithm.•The LDCRF model is applied to the domain of sketched symbol recognition.•A complete evaluation of the approach in the electric circuit domain is provided.•The achieved performance is compared with that of a baseline system.
论文关键词：Symbol recognition,Diagram understanding,Sketched diagrams,Discriminative models,Stroke clustering,Electric circuit diagrams
论文评审过程：Received 15 February 2012, Revised 11 September 2013, Accepted 19 September 2013, Available online 1 October 2013.
论文官网地址：https://doi.org/10.1016/j.patcog.2013.09.016
