Table 5.

Summary of detection results on the test set using a model trained for a two fruit classes.

Network/	Camera	overlap = 0.25						overlap = 0.5						overlap = 0.75
algorithm		Red			Green			Red			Green			Red			Green
		P	R	F1	P	R	F1	P	R	F1	P	R	F1	P	R	F1	P	R	F1
classical	Overall	0.54	0.81	0.65	0.62	0.72	0.67	0.44	0.66	0.53	0.47	0.54	0.50	0.27	0.40	0.32	0.23	0.26	0.24
segmentation
	Overall	0.84	0.93	0.88	0.9	0.94	0.92	0.81	0.9	0.85	0.88	0.91	0.89	0.66	0.73	0.69	0.73	0.75	0.74
	1	0.7	0.78	0.74	1	1	1	0.7	0.78	0.74	1	1	1	0.7	0.78	0.74	0.8	0.8	0.8
R50	2	0.88	0.93	0.9	0.81	0.97	0.88	0.86	0.91	0.88	0.76	0.91	0.83	0.68	0.72	0.7	0.63	0.75	0.68
	3	0.82	0.94	0.88	0.95	0.98	0.96	0.77	0.89	0.83	0.93	0.96	0.94	0.66	0.76	0.71	0.8	0.83	0.81
	4	0.67	1	0.8	0.86	0.79	0.82	0.67	1	0.8	0.81	0.74	0.77	0.33	0.5	0.4	0.56	0.51	0.53
	Overall	0.88	0.93	0.9	0.92	0.91	0.91	0.85	0.9	0.87	0.89	0.88	0.88	0.7	0.74	0.72	0.74	0.73	0.73
	1	1	0.78	0.88	1	1	1	1	0.78	0.88	1	1	1	1	0.78	0.88	0.8	0.8	0.8
R50+PP	2	0.92	0.93	0.92	0.84	0.94	0.89	0.9	0.91	0.9	0.78	0.88	0.83	0.72	0.72	0.72	0.66	0.74	0.7
	3	0.82	0.94	0.88	0.96	0.98	0.97	0.77	0.89	0.83	0.95	0.96	0.95	0.66	0.76	0.71	0.82	0.83	0.82
	4	0.75	1	0.86	0.85	0.64	0.73	0.75	1	0.86	0.79	0.6	0.68	0.5	0.67	0.57	0.55	0.41	0.47
	Overall	0.82	0.91	0.86	0.89	0.95	0.92	0.79	0.89	0.84	0.86	0.91	0.88	0.66	0.74	0.7	0.7	0.75	0.72
	1	0.64	1	0.78	0.71	1	0.83	0.57	0.89	0.69	0.71	1	0.83	0.57	0.89	0.69	0.71	1	0.83
R101	2	0.84	0.9	0.87	0.79	0.97	0.87	0.83	0.89	0.86	0.74	0.91	0.82	0.67	0.71	0.69	0.6	0.74	0.66
	3	0.83	0.92	0.87	0.94	0.98	0.96	0.78	0.87	0.82	0.92	0.96	0.94	0.7	0.77	0.73	0.78	0.82	0.8
	4	0.67	1	0.8	0.85	0.81	0.83	0.67	1	0.8	0.79	0.76	0.77	0.5	0.75	0.6	0.52	0.5	0.51
	Overall	0.88	0.91	0.89	0.91	0.91	0.91	0.86	0.89	0.87	0.88	0.88	0.88	0.73	0.75	0.74	0.72	0.73	0.72
	1	1	1	1	1	1	1	0.89	0.89	0.89	1	1	1	0.89	0.89	0.89	1	1	1
R101+PP	2	0.91	0.89	0.9	0.81	0.92	0.86	0.91	0.89	0.9	0.77	0.88	0.82	0.73	0.71	0.72	0.64	0.72	0.68
	3	0.84	0.92	0.88	0.96	0.98	0.97	0.79	0.87	0.83	0.94	0.96	0.95	0.71	0.77	0.74	0.8	0.82	0.81
	4	0.75	1	0.86	0.82	0.66	0.73	0.75	1	0.86	0.77	0.61	0.68	0.75	1	0.86	0.52	0.41	0.46
	Overall	0.94	0.93	0.93	0.93	0.95	0.94	0.93	0.93	0.93	0.91	0.93	0.92	0.82	0.82	0.82	0.75	0.77	0.76
	1	1	0.89	0.94	0.83	1	0.91	1	0.89	0.94	0.83	1	0.91	1	0.89	0.94	0.83	1	0.91
X101	2	0.94	0.91	0.92	0.85	0.97	0.91	0.93	0.9	0.91	0.81	0.92	0.86	0.84	0.81	0.82	0.62	0.71	0.66
	3	0.92	0.97	0.94	0.96	0.98	0.97	0.92	0.97	0.94	0.94	0.96	0.95	0.8	0.84	0.82	0.82	0.84	0.83
	4	1	1	1	0.92	0.83	0.87	1	1	1	0.89	0.8	0.84	0.5	0.5	0.5	0.65	0.59	0.62
	Overall	0.95	0.93	0.94	0.94	0.91	0.92	0.94	0.92	0.93	0.91	0.89	0.9	0.84	0.82	0.83	0.77	0.75	0.76
	1	1	0.89	0.94	1	1	1	1	0.89	0.94	1	1	1	1	0.89	0.94	1	1	1
X101+PP	2	0.96	0.9	0.93	0.86	0.94	0.9	0.95	0.89	0.92	0.82	0.89	0.85	0.86	0.81	0.83	0.63	0.69	0.66
	3	0.92	0.97	0.94	0.97	0.98	0.97	0.92	0.97	0.94	0.95	0.96	0.95	0.8	0.84	0.82	0.83	0.84	0.83
	4	1	1	1	0.92	0.67	0.78	1	1	1	0.88	0.64	0.74	0.67	0.67	0.67	0.67	0.49	0.57

Each row corresponds to inference results with one method/architecture. When depth post-processing is used, it is indicated by +PP.