Table 2.

Farm 1: Performance of three deep learning models considering all the data modalities using four different performance metrics, mean absolute error (MAE), root mean squared error (RMSE), mean absolute percentage error (MAPE), and R-squared ($R^2$). The MAE and RMSE errors can be interpreted as average errors in kilograms between the predicted and actual cattle weights, while the MAPE and R-Squared metrics represent the accuracy and goodness-of-fit of the models and do not have a direct interpretation in kilograms.

Model	Data Modalities	MAE (Kg)	RMSE (Kg)	MAPE (%)	R-Squared (${\mathit{R}}^2$)
INC	RGB	16.80	24.18	4.24	0.94
	DP	25.35	32.96	6.18	0.88
	RGBD	17.93	26.43	4.58	0.93
	FS	20.00	29.23	5.12	0.91
	FSD	17.91	24.39	4.58	0.90
MOB	RGB	16.24	22.74	4.38	0.95
	DP	23.09	31.86	6.03	0.89
	RGBD	17.56	23.78	4.56	0.94
	FS	20.54	29.47	5.08	0.91
	FSD	17.20	24.12	4.31	0.90
EFF	RGB	14.35	19.53	3.99	0.96
	DP	24.32	30.87	6.60	0.90
	RGBD	16.32	20.94	4.29	0.95
	FS	16.67	24.09	4.43	0.94
	FSD	16.48	20.59	4.58	0.93