Detection and analysis of wheat spikes using Convolutional Neural Networks

Table 2 Evaluation and validation of spike detection using the GSYC image model applied to the GSYC image data set

Image no	GT-count	Detected	TP	FP	FN	Precision	mAP	Accuracy	F1-score
Test_001.jpg	73	71	70	1	3	0.98	0.7289	96%	0.97
Test_012.jpg	75	68	68	0	7	1.00	0.6002	91%	0.95
Test_025.jpg	87	85	84	1	3	0.98	0.7324	96%	0.97
Test_032.jpg	80	76	76	0	4	1.00	0.7286	95%	0.97
Test_118.jpg	76	73	70	3	6	0.95	0.6126	92%	0.93
Test_141.jpg	66	61	58	3	8	0.95	0.5835	88%	0.91
Test_185.jpg	69	68	65	3	4	0.95	0.7105	94%	0.94
Test_199.jpg	72	69	68	1	4	0.98	0.7184	94%	0.96
Test_220.jpg	80	79	76	3	4	0.96	0.7229	95%	0.95
Test_242.jpg	70	64	63	1	7	0.90	0.5926	90%	0.94
Test_254.jpg	83	77	76	1	7	0.98	0.6085	91%	0.95
Test_320.jpg	80	77	74	3	6	0.96	0.6213	92%	0.94
Test_383.jpg	87	84	78	6	9	0.92	0.5947	90%	0.91
Test_399.jpg	80	78	77	1	3	0.98	0.7301	96%	0.97
Test_417.jpg	96	93	89	4	7	0.95	0.6573	93%	0.94
Test_421.jpg	71	73	70	3	1	0.95	0.7552	98%	0.97
Test_422.jpg	82	79	78	1	4	0.98	0.7211	95%	0.96
Test_432.jpg	85	81	79	2	6	0.97	0.6502	93%	0.95
Test_437.jpg	70	64	62	2	8	0.96	0.5924	88%	0.92
Test_480.jpg	88	84	82	2	6	0.97	0.6441	93%	0.95
Total	1570	1504	1463	41	107	−	−	−	−
Average	−	−	−	−	−	0.97	0.6653	93.4%	0.95
Standard dev.	7.82	8.17	7.86	1.46	1.11	0.02	0.06	0.03	0.02

ISSN: 1746-4811