Hi, This is Imtiaz

I'm a machine learning enthusiast/advocate in my first year of PhD with Dr. Richard Baraniuk at Rice University. Previously, I worked as a Research Assistant at the mHealth Research Group, Department of Biomedical Engineering, Bangladesh University of Engineering and Technology (BUET). I have completed my Bachelor's in Electrical and Electronic Engineering from BUET in September 2017. My research interests include Computer Vision, Biomedical Signal Processing, Domain Adaptation, Low Resource Speech Recognition and Music Information Retrieval.

Rashed, Samiul and I started working on Bengali.AI from December 2017 and have been leading this platform ever since. In my free time I love playing Sitar, and quite recently I have started self learning (read failing at) piano.

Google Scholar


My research mainly involves injecting signal processing intuition into Convolutional Neural Networks to adapt them for low resource tasks. A key problem for biomedical engineering tasks is the scarcity and domain dependency of data. Removal of domain dependency is key to widespread deployment of AI algorithms, namely, in mobile healthcare applications. My research is currently focused on conditioning CNNs so that the features learned by them lie between handcrafted features and generic CNN features, to ensure better generalization and less dependency on the domains/sensors.

  • Domain Adaptation

    feat tsnedeep tsne

    Biomedical data is at the mercy of the acquisition device. A model trained on data from one device tends to perform below par on data from a different one. This is a common problem when the volume of data is low and data from different devices are not available. I’m experimenting an adversarial domain adaptation method where a conditional adversarial discriminator forces the model to derive domain agnostic features.

  • Interpretable Machine Learning

    gradCAMlearned coeff

    Interpretability of machine learning algorithms, especially, neural networks, has been of utmost interest to researchers in recent years. My research explores the interpretability of 1DCNNs from a Finite Impulse Response/Signal Processing perspective.

  • mHealth and Pervasive Application

    HS AppPatient

    Mobile Healthcare (mHealth for short) deals with the application of healthcare systems in setups where the availability of physicians is scarce. Cardiovascular diseases are one of the major causes of death in Bangladesh, mostly due to the scarcity of trained cardiologists who can diagnose a cardiac anomaly at an early stage. My work incorporates deploying deep learning models for pre-screening of patients for cardiac abnormality.

  • Computer Vision for Assistive Technologies


    Embedded Application of Computer Vision Algorithms have wide potential in assistive technologies. Two key challenges remain vital for the success and outreach of such technologies: real-time algorithms and the price tag. My research incorporated the use of cheap webcams with a real-time embedded application. It also utilized Head Related Transfer Functions to modulate stereo sounds for blind assistance.

  • Music Information Retrieval


    Music Information Retrieval is the field of research that enables your Shazam, that powers your Spotify recommendations. I was introduced to it through the IEEE Signal Processing Cup 2017, and have fallen in love with it ever since. My interests lie in cover song identification and music beat tracking.


  • A. I. Humayun, S. Ghaffarzadegan, Z. Feng, T. Hasan, “Towards Stethoscope Invariant Heart Sound Abnormality Detection using Learnable Filterbanks”, IEEE Trans. BME, 2019 (Submitted).

  • A. I. Humayun, A. S. Sushmit, T. Hasan and M. I. H. Bhuiyan, “End-to-end Sleep Staging with Raw Single Channel EEG using Deep Residual ConvNets”IEEE BHI, Chicago, May, 2019.

  • A. S. Sushmit, S. U. Zaman, A. I. Humayun, T. Hasan and M. I. H. Bhuiyan, “X-Ray Image Compression Using Convolutional Recurrent Neural Networks”IEEE BHI, Chicago, May, 2019.

  • A. I. Humayun, M. T. Khan, S. Ghaffarzadegan, Z. Feng and T. Hasan, “An Ensemble of Transfer, Semi-supervised and Supervised Learning Methods for Pathological Heart Sound Classification”INTERSPEECH, India, September, 2018.

  • A. I. Humayun, S. Ghaffarzadegan, Z. Feng, T. Hasan, “Learning Front-end Filter-bank Parameters using Convolutional Neural Networks for Abnormal Heart Sound Detection”, IEEE EMBC, Hawaii, July, 2018

  • S. Alam, T. Reasat, R. Mohammad Doha, A. I. Humayun, NumtaDB – Assembled Bengali Handwritten Digits”, arXiv:1806.02452, 2018

  • S. A. Kamran, A. I. Humayun, S. Alam, R. Doha, M. Mandal, R. Tahsin, F. Rahman, “AI Learns to Recognize Bengali Handwritten Digits : Bengali.AI Computer Vision Challenge 2018”,  arXiv:1810.04452, 2018

  • I. A. Hussaini, A. I. Humayun, S. I. Foysal, S. Alam, A. Masud, A. Mahmud, R. Islam, N. Ibtehaz, S. U. Zaman, R. Hyder, S. S. Chowdhury, and M. A. Haque, “Predictive Real-time Beat Tracking from Music for Embedded Application”, IEEE MIPR, Miami, April, 2018.

  • A. I. Humayun, What Is the Future of Signal Processing?”, IEEE Signal Processing Magazine, Nov 2017, Page 14, Column 2.

Awards And Honors

ISCA Grant for Students and Young Scientists

September, 2018

Travel Grant for INTERSPEECH 2018

Honorable Mention, IEEE Signal Processing Cup 2017

July, 2017


Featured on BBC Media Action

February, 2017


2nd Place, IIT Techfest Innovation Challenge

December, 2016

Young Innovator of the Year 2016 from Bangladesh

November, 2016

Falling Walls Lab, 2016

Best Project, ICEEICT 2016

September, 2016

Regional Finalist for San Francisco, Hult Prize 2016

February, 2016

University of New South Wales (UNSW) gold medal

December, 2011

Gold medal for topping the International Assessment of Schools in Science and Mathematics