Method, software, and device for training an alarm system to classify audio of an event

1. A method for training an alarm system to classify audio of an event, wherein the alarm system is connected to a neural network trained to classify audio as an event type, the method comprising:

detecting an event;

transmitting audio recorded during a first period of time comprising a time of the event to an external unit;

receiving data from the external unit indicating a sub-period of time of the audio and data indicating an event type of the indicated sub-period of time of the audio; and

re-training the neural network while the alarm system is in use by inputting a sub-period of the audio corresponding to the indicated sub-period of time of the audio and using the indicated event type as a correct classification of the sub-period of the audio.

2. The method according to claim 1 , further comprising:

recording video during the first period of time; and

transmitting the video to the external unit.

3. The method according to claim 1 , further comprising: analysing the audio by inputting the audio into the neural network, wherein the neural network outputs a suggested event type of the recorded audio; and

wherein the step of transmitting the audio further comprises transmitting data indicating the suggested event type of the audio.

4. The method according to claim 3 , wherein the received data indicating an event type comprises a verification of the suggested event type of the audio.

5. The method according to claim 1 , further comprising:

recording audio during a second period of time comprising the first period of time; and

analysing the recorded audio to identify audio indicating an event, wherein the audio recorded during the first period of time comprises the audio indicating an event.

6. The method according to claim 5 , wherein

analysing the recorded audio further comprises identifying audio indicating a conversation, and wherein the audio recorded during the first period of time excludes the audio indicating a conversation.

7. The method according to claim 1 , wherein transmitting the audio to the external unit further comprises transmitting coordinates of the alarm system.

8. The method according to claim 1 , wherein the neural network comprises n>1 layers, and the step of re-training the neural network comprises re-training only the last k>0 layers without changing the initial n−k layers, wherein n>k.

9. The method according to claim 1 , further comprising: communicating the data received from the external unit to another alarm system.

10. The method according to claim 9 , wherein: communicating the data comprises analysing the data indicating an event type and communicating the data received from the external unit to another alarm system that is determined to be interested by the indicated event type.

11. A non-transitory computer-readable storage medium with instructions adapted to carry out the method of claim 1 when executed by a device having processing capability.

12. A device for classifying audio of an event detected by an alarm system, wherein the device is connected to a neural network trained to classify audio as an event type, the device comprising:

a transmitter configured to transmit audio recorded during a first period of time comprising a time of the event to an external unit;

a receiver configured to receive data from the external unit indicating a sub-period of time of the audio and data indicating an event type of the indicated sub-period of time of the audio; and

a processing circuit configured to re-train the neural network while the alarm system is in use by inputting a sub-period of the audio corresponding to the indicated sub-period of time of the audio and using the indicated event type as a correct classification of the sub-period of the audio.

13. The device according to claim 12 , wherein the device is further configured to analyse the recorded audio by inputting it into the neural network, wherein the neural network outputs a suggested event type of the recorded audio;

wherein the transmitter is further configured to transmit data indicating the suggested event type of the audio; and

wherein the second receiver configured to receive data from the external unit is further configured to receive a verification of the suggested event type of the audio.

14. The device according to claim 12 , wherein the device is further configured to:

receive audio recorded during a second period of time comprising the first period of time; and

analyse the recorded audio to identify audio indicating an event, wherein the audio recorded during the first period of time comprises the audio indicating an event.

15. A system comprising at least two devices according to claim 12 , each device further being configured to communicate the data received from the external unit to each other.