Electronic apparatus and controlling method thereof

An electronic apparatus comprising:

• a display;
• a microphone;
• a memory storing instructions;
• and
• a processor configured to execute the instructions to:
• control the display to display a content,
• control the microphone to receive an utterance of a user while the content is displayed on the display,
• identify a dedicated language model based on the content displayed on the display,
• recognize, through an automatic speech recognition (ASR) function of the electronic apparatus, the received utterance to identify candidate texts of the recognized utterance,
• identify a similarity between the recognized utterance and each of the identified candidate texts,
• identify, based on the identified dedicated language model and a predetermined threshold value, a suitability of a predetermined number of candidate texts with a high identified similarity, among the candidate texts, by calculating a total sum of perplexity values of the predetermined number of candidate texts, the predetermined number being two or more,
• based on the identified suitability of the predetermined number of candidate texts being outside a predetermined suitability range, ignore the recognized utterance, and
• based on the identified suitability of the predetermined number of candidate texts being in the predetermined suitability range, identify a candidate text having a highest suitability, among the candidate texts, as the recognized utterance and perform an operation corresponding to the recognized utterance.

A non-transitory computer readable medium for storing computer readable program code or instructions which are executable by one or more processors to perform a method of controlling an electronic apparatus, the method comprising:

• displaying, by the electronic apparatus, a content on a display of the electronic apparatus;
• identifying, by the electronic apparatus, a dedicated language model based on the content displayed on the display of the electronic apparatus;
• receiving, by the electronic apparatus, an utterance of a user while the content is displayed on the display of the electronic apparatus;
• recognizing, by the electronic apparatus, the received utterance through an automatic speech recognition (ASR) function of the electronic apparatus and identifying, by the electronic apparatus, candidate texts of the recognized utterance;
• identifying, by the electronic apparatus, a similarity between the recognized utterance and each of the identified candidate texts;
• identifying, by the electronic apparatus, based on the identified dedicated language model and a predetermined threshold value, a suitability of a predetermined number of candidate texts with a high identified similarity, among the candidate texts, by calculating a total sum of perplexity values of the predetermined number of candidate texts, the predetermined number being two or more;
• based on the identified suitability of the predetermined number of candidate texts being outside a predetermined suitability range, ignoring, by the electronic apparatus, the recognized utterance;
• and
• based on the identified suitability of the predetermined number of candidate texts being in the predetermined suitability range, identifying, by the electronic apparatus, a candidate text having a highest suitability, among the candidate texts, as the recognized utterance, and performing, by the electronic apparatus, an operation corresponding to the recognized utterance.