Skip to content

AI-Powered Animal Sound Translator & Taxonomy Identifier. Translate animal sounds into human-understandable behavior and identify species from images using AI models, audio processing, and open-data APIs.

Notifications You must be signed in to change notification settings

Aashritha-2005/echoAnimal

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

4 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

🐾 Animal Sound Translator & Taxonomy Identifier

Welcome to the AI-Powered Animal Translator & Taxonomy Identifier, a dual-module project that brings together AI, audio/image processing, and open data APIs to help humans better understand and identify animals — through both sound and sight.

🚀 Overview

Animals communicate in many fascinating ways — through vocalizations and visual cues. Our project is split into two main modules:

  1. 🐾 Animal Sound Translator — Analyzes animal audio and generates behavior-based translations.
  2. 🧬 Animal Image Identifier — Classifies animals from images and retrieves their full taxonomy.

Features

  • Real-time Audio Analysis: Upload audio files and get instant analysis

  • AI-Powered Classification: Uses pre-trained models from Hugging Face to identify animals

  • Visual Analysis:

    • Audio waveform visualization
    • Detailed spectrograms
    • Confidence level charts
  • Smart Translation: Converts animal sounds to human language based on:

    • Audio characteristics (pitch, energy, frequency)
    • Animal behavior patterns
    • Confidence levels
  • Comprehensive Metrics:

    • Audio features (MFCC, spectral analysis)
    • Pitch analysis and variation
    • Energy levels and patterns

    🐯 Animal Image Identifier

  • Image classification using ResNet-50 via Hugging Face

  • Taxonomy retrieval using Wikipedia & Wikidata APIs

  • Full hierarchy extraction from species to kingdom

  • Clean UI built in Streamlit

Supported Animals

  • Dogs (barking, whining, howling)
  • Cats (meowing, purring)
  • Birds (chirping, singing)
  • Cows (mooing)
  • Horses (neighing)
  • Pigs (oinking)
  • Sheep (bleating)
  • And many more!

🛠️ Technologies Used

Category Tools/Tech
Language Python 3.8+
Web Framework Streamlit
AI Models Hugging Face (ResNet-50, Audio)
Signal Processing librosa
APIs Wikipedia API, Wikidata API

Installation

  1. Install Python 3.8 or higher
  2. Install required packages:
python -m pip install -r requirements.txt

Usage

  1. Run the application:
streamlit run echoanimal.py
  1. Open your browser and navigate to the provided URL (usually http://localhost:8501)

  2. Upload an audio file containing animal sounds

  3. View the analysis results:

    • Animal identification with confidence scores
    • Audio visualizations (waveform and spectrogram)
    • Human translation of the animal communication
    • Detailed audio insights and metrics

Technical Details

AI Models

  • Uses pre-trained audio classification models from Hugging Face
  • Specifically optimized for animal sound recognition
  • Confidence scoring with uncertainty handling

Audio Processing

  • librosa: Advanced audio analysis and feature extraction
  • MFCC Analysis: Mel-frequency cepstral coefficients for sound characterization
  • Spectral Features: Centroid, bandwidth, rolloff analysis
  • Pitch Detection: Fundamental frequency analysis
  • Energy Analysis: RMS energy and zero-crossing rate

Translation Logic

  • Based on scientific animal behavior research
  • Considers audio characteristics:
    • Pitch (high/low frequency content)
    • Energy (loud/soft intensity)
    • Temporal patterns (rapid/slow variations)
  • Contextual interpretation based on animal species
  • Confidence-weighted translations

File Formats Supported

  • WAV
  • MP3
  • OGG
  • FLAC
  • M4A

Technical Approach — Taxonomy Identification 🧠

1.Image Classification:Hugging Face pipeline with microsoft/resnet-50

2.Wikipedia API:Correct species title retrieval

3.Wikidata API:Recursive search to fetch complete taxonomy levels

4.Automated extraction up to Kingdom level

Limitations

  • Translation is based on general animal behavior patterns, not actual animal language
  • Accuracy depends on audio quality and clarity
  • Best results with clear, isolated animal sounds
  • Some exotic animals may not be recognized

Future Enhancements

  • Real-time microphone input
  • Multi-animal sound separation
  • Emotional state detection
  • Extended animal database
  • Mobile app version

Contributing

Feel free to contribute by:

  • Adding support for more animals
  • Improving translation accuracy
  • Enhancing the user interface
  • Adding new audio analysis features

🧠 Features

🔊 Animal Sound Translator

  • Real-time audio analysis using librosa & Hugging Face audio models
  • Feature extraction: MFCCs, pitch, energy, spectral features
  • Behavior translation: Maps sound patterns to behaviors using science-backed logic
  • Visualizations: Audio waveform, spectrograms, confidence scores

About

AI-Powered Animal Sound Translator & Taxonomy Identifier. Translate animal sounds into human-understandable behavior and identify species from images using AI models, audio processing, and open-data APIs.

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages