Convmelspec: Convertible Melspectrograms via 1D Convolutions

Easier audio-based machine learning with TensorFlow.

A neural network framework for researchers studying acoustic communication

Environmental sound classification with Convolutional neural networks and the UrbanSound8K dataset.

Deep learning using CNN for Mandarin Chinese tone classification

NTU RGB+D Dataset Action Recognition with GNNs and CNNs

Streamline your ecoacoustic analysis with LEAVES, offering advanced tools for large-scale soundscape annotation and visualization. Join researchers and citizen scientists using LEAVES to analyze complex soundscapes faster and more accurately.

Find gravitational wave signals from binary black hole collisions.

Using deep learning techniques like 1D and 2D CNNs, LSTM to detect damage in a structure with hinges/joints after an earthquake.

Music timbre transfer

Tackle accent classification and conversion using audio data, leveraging MFCCs and spectrograms. Models differentiate accents and convert audio between accents

NoiseCapture 2 Multi Platform

Classifying Radio signal coming from space

Code and material relevant to the paper "Spectrogrand: Computational Creativity Driven Audiovisuals' Generation From Text Prompts"

TinyML project. This system monitors your room or surrounding with an onboard microphone of Arduino nano BLE sense. Still Under Developement

LOFAR System Health Management

A Python script that transforms images into audio

Bachelor Final Year Project exploring real-time speech denoising using machine learning. Compares classical methods (SS, WF, MMSE-LSA) with 5 deep models on spectrogram data, highlighting Conv-TasNet’s effectiveness. Features dataset bucketing, OOM mitigation, and batch evaluation.

Analysis of human behavioral and neural variability during naturalistic arm movements. Replicates the findings in our preprint: https://www.biorxiv.org/content/10.1101/2020.04.17.047357v2

Python package for reproducible feature extraction from spectrograms with custom pre-trained encoders

Data Preprocessing and Exploratory Data Analysis project, CNN model

Interactive app for pre-annotation of spectrogram images aided by deep-leaning based clustering

Build a Neural Network to identify and classify emotion Real-time Emotion Detection using the tone of their voice. Restrictive to English language (American accent)

Implementation in Python of a tool to automatically classify speech segments according to intonation system of Cuba.

The code implements the Deep CNN model described in Salamon and Bello's paper for Environmental Sound Classification on Urbansound8k dataset

Repository for COVID-19 screening project. Involves audio processing and some CV.

Different Signal Processing Tasks

Analyzes aquatic audio to distinguish shrimp feeding/snapping from insect noise using signal processing for clearer underwater sound analysis.

Vision Transformer vs CNN for LIGO gravitational wave glitch classification — class-dependent architecture preferences and CW search implications

Find how similar your voice is to Taylor Swift (WIP) ✨