Method: Multi_scale_v2 - Task 2 - Script identification - ICDAR2017 Competition on Multi-lingual scene text detection and script identification

method: Multi_scale_v22019-09-27

Authors: Wuheng Xu, Changxu Cheng, Bohan Li

Description: Area block feature information using images on multiple scales.This model has 4 scales and 8 branches.We used three training sets(mlt17, mlt19, mlt19val).

Confusion Matrix

		Detection
		Arabic	Latin	Chinese	Japanese	Korean	Bangla	Symbols	Mixed	None
GT	Arabic	4708	340	22	35	18	7	12	0	0
	Latin	115	59199	215	459	318	81	150	0	0
	Chinese	2	325	3689	679	34	8	13	0	0
	Japanese	13	1821	1060	5099	131	25	8	0	0
	Korean	25	1682	349	372	10524	32	8	0	0
	Bangla	6	204	20	35	11	2268	1	0	0
	Symbols	39	821	12	71	39	4	2510	0	0
	Mixed	0	0	0	0	0	0	0	0	0
	None	0	0	0	0	0	0	0	0	0