Table 5. Evaluating time-cost for automatic identity correction – comparing to results from idtracker.ai.

Timings consist of preprocessing time in TGrabs plus network training in TRex, which are shown separately as well as combined (ours (min), $N=5$). The time it takes to analyze videos strongly depends on the number of individuals and how many usable samples per individual the initial segment provides. The length of the video factors in as well, as does the stochasticity of the gradient descent (training). idtracker.ai timings ($N=3$) contain the whole tracking and training process from start to finish, using its terminal_mode (v3). Parameters have been manually adjusted per video and setting, to the best of our abilities, spending at most one hour per configuration. For videos 16 and 14, we had to set idtracker.ai to storing segmentation information on disk (as compared to in RAM) to prevent the program from being terminated for running out of memory.

Video	# ind.	Length	Sample	TGrabs (min)	TRex (min)	Ours (min)	idtracker.ai (min)
7	100	1min	1.61s	2.03 ± 0.02	74.62 ± 6.75	76.65	392.22 ± 119.43
8	59	10min	19.46s	9.28 ± 0.08	96.7 ± 4.45	105.98	495.82 ± 115.92
9	15	60min	33.81s	13.17 ± 0.12	101.5 ± 1.85	114.67	N/A
11	10	10min	12.31s	8.8 ± 0.12	21.42 ± 2.45	30.22	127.43 ± 57.02
12	10	10min	10.0s	8.65 ± 0.07	23.37 ± 3.83	32.02	82.28 ± 3.83
13	10	10min	36.91s	8.65 ± 0.047	12.47 ± 1.27	21.12	79.42 ± 4.52
10	10	10min	16.22s	4.43 ± 0.05	35.05 ± 1.45	39.48	N/A
14	8	195min	67.97s	109.97 ± 0.05	70.48 ± 3.67	180.45	707.0 ± 27.55
15	8	72min	79.36s	32.1 ± 0.42	30.77 ± 6.28	62.87	291.42 ± 16.83
16	8	198min	134.07s	133.1 ± 2.28	68.85 ± 13.12	201.95	1493.83 ± 27.75