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

method: ecn-based method2017-06-02

Authors: Yizhi Wang

Description: This is a ECN-based model. It uses a loss function that takes into account the global classification error for a group of N patches instead of looking only into a single image patch.
ECN: Gomez L, Nicolaou A, Karatzas D. Improving patch-based scene text script identification with ensembles of conjoined networks[J]. Pattern Recognition, 2017, 67:85-96.

Confusion Matrix

		Detection
		Arabic	Latin	Chinese	Japanese	Korean	Bangla	Symbols	Mixed	None
GT	Arabic	4425	520	44	90	41	9	13	0	0
	Latin	328	57571	617	924	797	169	131	0	0
	Chinese	11	440	3245	941	105	3	5	0	0
	Japanese	58	2667	1144	3945	308	23	12	0	0
	Korean	123	4104	469	802	7412	72	10	0	0
	Bangla	18	269	37	46	21	2154	0	0	0
	Symbols	61	1602	25	228	67	19	1494	0	0
	Mixed	0	0	0	0	0	0	0	0	0
	None	0	0	0	0	0	0	0	0	0