Abstract: Systems and methods for training models to predict dense correspondences across images such as human images. A model may be trained using synthetic training data created from one or more 3D computer models of a subject. In addition, one or more geodesic distances derived from the surfaces of one or more of the 3D models may be used to generate one or more loss values, which may in turn be used in modifying the model's parameters during training.

Abstract: Disclosed are a dual-motor multi-gear hybrid transmission system and a vehicle. The dual-motor multi-gear hybrid transmission system includes an engine, a first motor, a second motor, a first clutch, a second clutch, a first planet row, a second planet row, a first input shaft, a second input shaft, a third input shaft and a brake assembly. The first input shaft is connected to the engine through the first clutch. The second input shaft is connected to the engine through the second clutch, and the second input shaft is sleeved outside the first input shaft. The first motor is connected to the engine. The second motor is connected to the first planet row through the third input shaft. The brake assembly is configured to brake the first planet row and/or the second planet row.

Abstract: The present disclosure provide a method for recognizing a space frequency block code (SFBC), including: obtaining a cross-correlation function of a reception signal, and obtaining a Fourier transformation of the cross-correlation function; obtaining data containing labels by labeling a sample containing a real part and an imaginary part of the Fourier transformation, and inputting the data containing labels into an initial CNN-LSTM network for training until convergence is reached, to obtain a final CNN-LSTM network; and obtaining unlabeled data to be recognized, and inputting the data to be recognized into the final CNN-LSTM network to obtain a SFBC recognition result.