Critical Literature Review of Named Entity Recognition in Symbol Spotting

Abstract

Symbol spotting is a subset of image recognition that focuses on identifying objects in an image. Symbol spotting has received some research interest lately. Symbol spotting is beneficial for image analytics of engineering diagrams where the goal is to retrieve specific symbols. There are some works done to identify type of entity but very little for Name Entity Recognition (NER). NER is a machine learning feature that can help retrieval of more precise symbols. Semantics analysis of NER in image analytics researches are extensively reviewed. This paper critically compares the state-of-the-arts, merits and gaps in the works. Review reveals the need to focus on techniques to increase accuracy of symbol spotting.