A Review on Text Detection for Multi-Orientation Scene Images using Adaptive Clustering

Detection of text in scene images is an important requirement for several content-based multimedia applications. In the real world, text detection task is often challenging due to issues such as complex background and variation of text orientation, font, size, color and to detect and recognize text in a unified manner by searching for words from the image into text regions or individual characters. Text captured in natural scenes is most of the times with multiple orientations and perspective distortions while currently most research efforts only focus on horizontal or near horizontal scene text. To address this issues a novel approach unified distance metric learning framework is proposed based on adaptive hierarchical clustering, which can simultaneously learn weights similarity to adaptively combine different feature similarities and the clustering threshold to determine the number of clusters and then an effective multi-orientation scene text detection system, which constructs the text candidates by grouping characters based on this adaptive clustering.