| --- |
| dataset_info: |
| name: Animal Sound Classification Dataset |
| type: audio-classification |
| license: mit |
| annotations_creators: |
| - expert-generated |
| language: |
| - no-linguistic-content |
| task_categories: |
| - audio-classification |
| pretty_name: Animal Sound Classification |
| size_categories: 1K<n<10K |
| tags: |
| - animal-sounds |
| - audio |
| - sound-classification |
| - environmental-sounds |
| - MFCC |
| - open-dataset |
| - sound-recognition |
| features: |
| - name: filename |
| type: string |
| - name: mfcc_1 |
| type: float64 |
| - name: mfcc_2 |
| type: float64 |
| - name: mfcc_3 |
| type: float64 |
| - name: mfcc_4 |
| type: float64 |
| - name: mfcc_5 |
| type: float64 |
| - name: mfcc_6 |
| type: float64 |
| - name: mfcc_7 |
| type: float64 |
| - name: mfcc_8 |
| type: float64 |
| - name: mfcc_9 |
| type: float64 |
| - name: mfcc_10 |
| type: float64 |
| - name: mfcc_11 |
| type: float64 |
| - name: mfcc_12 |
| type: float64 |
| - name: mfcc_13 |
| type: float64 |
| splits: |
| - name: train |
| num_bytes: 114400 |
| num_examples: 1045 |
| creators: |
| - name: Muhammad Qasim |
| url: https://github.com/MuhammadQasim111 |
| license: mit |
| --- |
| |
| # 🐾 Animal Sound Classification Dataset |
|
|
| > **A meticulously handcrafted dataset of labeled animal sounds for Machine Learning & Audio Classification tasks.** |
| > **Built with love, precision, and open-source spirit.** |
|
|
| --- |
|
|
| ## 📖 Dataset Details |
|
|
| ### 📝 Dataset Description |
|
|
| The **Animal Sound Classification Dataset** contains curated audio clips of **dogs, cats, cows**, and more, extracted from longer recordings and meticulously trimmed to create clean, high-quality sound samples. Over a period of **two months**, I manually processed, trimmed, and labeled each audio file. I also prepared the dataset for ML pipelines by extracting **MFCC (Mel-Frequency Cepstral Coefficients)** features to ensure seamless integration for developers and researchers. |
|
|
| - **Curated by:** Muhammad Qasim |
| - **Funded by:** Self-initiated Open-Source Project |
| - **License:** MIT License |
| - **Language(s):** Non-linguistic (animal sounds) |
|
|
| --- |
|
|
| ## 🔗 Dataset Sources |
|
|
| - **Repository:** [Hugging Face Link](https://huggingface.co/datasets/MuhammadQASIM111/Animal_Sound_Classification) |
|
|
| --- |
|
|
| ## 🚀 Uses |
|
|
| ### ✅ Direct Use |
|
|
| - Audio classification model training. |
| - Sound recognition AI systems. |
| - Educational apps that teach animal sounds. |
| - Wildlife and livestock sound monitoring AI. |
|
|
| ### ❌ Out-of-Scope Use |
|
|
| - Speech Recognition tasks. |
| - Use in sensitive environments without proper augmentation. |
| - Misuse for deceptive simulations. |
|
|
| --- |
|
|
| ## 🗂️ Dataset Structure |
|
|
| ### Data Instances |
|
|
| | Field Name | Type | Description | |
| |------------|------|-------------| |
| | filename | string | Name of the audio file | |
| | mfcc_1 | float64 | First MFCC feature | |
| | mfcc_2 | float64 | Second MFCC feature | |
| | ... | ... | ... | |
| | mfcc_13 | float64 | Thirteenth MFCC feature | |
| |
| ### Data Fields |
| |
| - `filename`: Name of the audio file. |
| - `mfcc_1` to `mfcc_13`: Mel-frequency cepstral coefficients (MFCCs) extracted from the audio files. |
| |
| ### Data Splits |
| |
| | Split | Number of Examples | Total Size | |
| |-------|--------------------|------------| |
| | Train | 1045 | 114.4 KB | |
| |
| --- |
| |
| ## 📥 Dataset Creation |
| |
| ### Curation Rationale |
| |
| The dataset was created to facilitate research and development in the field of audio classification, particularly focusing on animal sounds. The goal is to provide a high-quality, ready-to-use dataset for machine learning practitioners and researchers. |
| |
| ### Source Data |
| |
| #### Initial Data Collection and Normalization |
| |
| - **Data Collection:** Audio clips were collected from various sources and manually trimmed to isolate individual animal sounds. |
| - **Annotations:** Each audio clip was labeled with the corresponding animal class. |
| - **Who are the annotators?** The annotations were generated by an expert. |
| |
| ### Personal and Sensitive Information |
| |
| The dataset does not contain any personal or sensitive information. |
| |
| --- |
| |
| ## 📜 Additional Information |
| |
| ### Dataset Curators |
| |
| Muhammad Qasim |
| |
| ### Licensing Information |
| |
| MIT License |
| |
| ### Citation Information |
| |
| |
