Contextual automated audio talkdown for remote guarding

A method comprising:

• in an electronic device:
• detecting a movement of an object in a physical area that is inside or in a vicinity of a built environment;
• receiving, based on the detected movement, a sequence of images of the physical area that include the object;
• detecting one or more physical activities of the object that are associated with a behavior of an intruder, based on application of a first Artificial Intelligence (AI) model on the sequence of images;
• generating information that includes a textual description of the detected one or more physical activities, wherein the generation of the information including the textual description is based on application of a second AI model on an output of the first AI model, and the second AI model is configured to translate one or more text labels for the detected one or more physical activities and one or more attributes of the object to a natural language description;
• generating an audio alert based on the generated information, wherein the generation of the audio alert comprises:
• converting the textual description to an audio message;
• selecting an audio template from a set of audio templates based on the detected one or more physical activities, wherein
• the selected audio template includes a set of audio slots, and
• the set of audio slots includes an object slot associated with a type of the object, an attribute slot associated with the one or more attributes of the object, and a warning slot associated with the detection of the one or more physical activities;
• and
• inserting the audio message in a corresponding audio slot of the set of audio slots of the selected audio template;
• and
• controlling a playback of the audio alert via an audio reproduction device that is installed in a vicinity of the built environment, wherein the playback includes a recitation of the textual description included in the generated information.

A non-transitory computer-readable storage medium having stored thereon, computer-executable instructions that when executed by an electronic device, causes the electronic device to execute operations, the operations comprise:

• detecting a movement of an object in a physical area that is inside or in a vicinity of a built environment;
• receiving, based on the detected movement, a sequence of images of the physical area that include the object;
• detecting one or more physical activities of the object that are associated with a behavior of an intruder, based on application of a first Artificial Intelligence (AI) model on the sequence of images;
• generating information that includes a textual description of the detected one or more physical activities, wherein the generation of the information including the textual description is based on application of a second AI model on an output of the first AI model, and the second AI model is configured to translate one or more text labels for the detected one or more physical activities and one or more attributes of the object to a natural language description;
• generating an audio alert based on the generated information, wherein the generation of the audio alert comprises:
• converting the textual description to an audio message;
• selecting an audio template from a set of audio templates based on the detected one or more physical activities, wherein
• the selected audio template includes a set of audio slots, and
• the set of audio slots includes an object slot associated with a type of the object, an attribute slot associated with the one or more attributes of the object, and a warning slot associated with the detection of the one or more physical activities;
• and
• inserting the audio message in a corresponding audio slot of the set of audio slots of the selected audio template;
• and
• controlling a playback of the audio alert via an audio reproduction device that is installed in a vicinity of the built environment, wherein the playback includes a recitation of the textual description included in the generated information.