Scribd
Upload
19 views8 pages

Speech Recognition Introduction

Speech recognition, also known as Automatic Speech Recognition (ASR), converts spoken language into text and is utilized in various applications such as voice assistants and automated customer service. The typical speech recognition pipeline includes audio input, preprocessing, feature extraction, and decoding, while challenges include accents, background noise, and real-time performance needs. Tools like Google Speech-to-Text API and Python libraries facilitate the implementation of speech recognition systems.

Uploaded by

prabu
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
19 views8 pages

Speech Recognition Introduction

Speech recognition, also known as Automatic Speech Recognition (ASR), converts spoken language into text and is utilized in various applications such as voice assistants and automated customer service. The typical speech recognition pipeline includes audio input, preprocessing, feature extraction, and decoding, while challenges include accents, background noise, and real-time performance needs. Tools like Google Speech-to-Text API and Python libraries facilitate the implementation of speech recognition systems.

Uploaded by

prabu
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd

Introduction to Speech Recognition

• • Converts spoken language into text


• • Also called Automatic Speech Recognition
(ASR)
• • Bridges the gap between human speech and
computers
Real-World Use Cases
• • Voice assistants: Siri, Alexa, Google Assistant
• • Speech-to-text transcription (e.g., YouTube
captions)
• • Voice search and commands
• • Automated customer service
• • Accessibility for differently-abled
Typical Speech Recognition
Pipeline
• 1. Audio Input
• 2. Preprocessing (noise removal,
normalization)
• 3. Feature Extraction (MFCCs, spectrograms)
• 4. Acoustic Model
• 5. Language Model
• 6. Decoder
Why Speech Recognition is Hard
• • Accents and dialects
• • Background noise
• • Speaker variations
• • Homophones (e.g., write vs right)
• • Real-time performance needs
Components of a Speech
Recognition System
• • Acoustic Model: Maps audio to phonemes
• • Language Model: Predicts word sequence
• • Lexicon: Phonemes to words
• • Decoder: Combines models for transcription
Tools for Speech Recognition
• • Google Speech-to-Text API
• • CMU Sphinx (PocketSphinx)
• • Mozilla DeepSpeech
• • Facebook Wav2Vec 2.0
• • Python’s SpeechRecognition library
Example in Python
• import speech_recognition as sr

• r = sr.Recognizer()
• with sr.Microphone() as source:
• print('Say something...')
• audio = r.listen(source)

• try:
• print('You said: ' +
Speech Recognition – Key Points
• • Converts voice to text using ML
• • Involves acoustic and language modeling
• • Many real-world applications
• • Python and APIs make it accessible

You might also like