Orientation predicting method, virtual reality headset and non-transitory computer-readable medium

1. An orientation predicting method, adapted to a virtual reality headset, comprising:

obtaining an orientation training data and an adjusted orientation data, wherein the adjusted orientation data is obtained by cutting a data segment off from the orientation training data, wherein the data segment corresponds to a time interval determined by an application latency;

training an initial neural network model based on the orientation training data and the adjusted orientation data for obtaining a trained neural network model corresponding to the time interval;

retrieving a real-time orientation data by an orientation sensor of the virtual reality headset; and

inputting the real-time orientation data to the trained neural network model to output a predicted orientation data,

wherein the trained neural network model is one of a plurality of candidate neural network models, with said candidate neural network models corresponding to different time intervals respectively, and wherein before retrieving the real-time orientation data by the orientation sensor of the virtual reality headset, the method further comprises:

estimating a latency of an application run by the virtual reality headset; and

selecting the trained neural network model from the plurality of candidate neural network models according to the latency, with the latency closer to the time interval corresponding to the trained neural network model than to the time intervals corresponding to the others of the candidate neural network models.

2. The method according to claim 1 , wherein the initial neural network model comprises a one-dimensional convolutional neural network.

3. A virtual reality headset, comprising:

an orientation sensor, retrieving a real-time orientation data of the virtual reality headset;

a processor, inputting the real-time orientation data to a trained neural network model of the processor for obtaining a predicted orientation data, wherein the trained neural network model is obtained by training an initial neural network model based on an orientation training data and an adjusted orientation data, wherein the adjusted orientation data is obtained by cutting a data segment off from the orientation training data, wherein the data segment corresponds to a time interval determined by an application latency;

a screen, displaying a predicted image according to the predicted orientation data; and

a memory storing a plurality of candidate neural network models including the trained neural network model, wherein before inputting the real-time orientation data to the trained neural network model, the processor further estimates a latency of an application run by the virtual reality headset, and selects the trained neural network model from the plurality of candidate neural network models, with the latency closer to the time interval corresponding to the trained neural network model than to the time intervals corresponding to the others of the candidate neural network models.

4. The virtual reality headset according to claim 3 , wherein the initial neural network model comprises a one-dimensional convolutional neural network.

5. A non-transitory computer-readable medium, storing an executable instruction which, when executed, causes a virtual reality headset to perform a method comprising:

retrieving a real-time orientation data by an orientation sensor and inputting the real-time orientation data to a trained neural network model to output a predicted orientation data, wherein the trained neural network model is obtained by training an initial neural network model based on an orientation training data and an adjusted orientation data, wherein the adjusted orientation data is obtained by cutting a data segment off from the orientation training data, wherein the data segment corresponds to a time interval determined by an application latency,

wherein the trained neural network model is one of a plurality of candidate neural network models, with said candidate neural network models corresponding to different time intervals respectively, and wherein before receiving the real-time orientation data from the orientation sensor, further comprises:

estimating a latency of an application run by the virtual reality headset; and

selecting the trained neural network model from the plurality of candidate neural network models according to the latency, with the latency closer to the time interval corresponding to the trained neural network model than to the time intervals corresponding to the others of the candidate neural network models.

6. The non-transitory computer-readable medium according to claim 5 , wherein the neural network model comprises a one-dimensional convolutional neural network.