Shahed Ahmed

I am a first year PhD student at Purdue ECE, advised by Prof. Joseph Makin. My research interests are broadly in the areas of signal processing, machine learning, and deep learning. I am currently working on developing deep learning models that can decode human speech activity from intercranial neural signal recordings.

Prior to joining Purdue, I completed my undergraduate and master's degrees from the department of Electrical and Electronic Engineering (EEE), BUET in 2021 and 2023, respectively. I have also taught at the department of EEE, BUET as a lecturer for three years.

Outside of work, I like to travel, read books, and play chess in my free time.

Email / CV / Google Scholar / LinkedIn / ResearchGate

News and Updates

08.19.24: Started my PhD journey at Purdue University!!
08.09.24: Arrived at West Lafayette, Indiana!
12.24.23: Research article on image segmentation got accepted in Computer Vision and Image Understanding journal
09.16.23: Successfully defended my Master's thesis!
08.03.23: Recieved my CPD certified certificate of particiaption for Oxford Machine Learning Summer School, 2023
07.16.23: Finished the MLxHealth track from Oxford Machine Learning Summer School, 2023. Great Experience!
06.22.23: Research paper on medical image segmentation got accepted in Biomedical Signal Processing and Control journal

Publications

Representative papers are highlighted.

	Twin-SegNet: Dynamically coupled complementary segmentation networks for generalized medical image segmentation Shahed Ahmed, Md.Kamrul Hasan Computer Vision and Image Understanding, 2024 We propose a dual-stream segmentation model, Twin-SegNet that performs both foreground and background tissue segmentation, and unifies them through a dynamic partial channel recalibration technique.
	COMA-Net: Towards generalized medical image segmentation using complementary attention guided bipolar refinement modules Shahed Ahmed, Md.Kamrul Hasan Biomedical Signal Processing and Control, 2023 We propose COMA-Net for achieving generalized medical image segmentation. We show that increasing the strength of foreground tissue w.r.t. the background at the feature space is beneficial to overall segmentation performance.
	S²C-DeLeNet: A parameter transfer based segmentation-classification integration for detecting skin cancer lesions from dermoscopic images Jahin Alam, Mir Sayeed Mohammad, Md Adnan Faisal Hossain, Ishtiaque Ahmed Showmik, Munshi Sanowar Raihan, Shahed Ahmed, Talha Ibn Mahmud Computers in Biology and Medicine, 2022 A novel framework is proposed that unifies classification and segmentation strategies for efficient skin lesion segmentation.
	CapNet: A Deep Learning-based Framework for Estimation of Capnograph Signal from PPG Shahed Ahmed, Md. Tariqul Islam, Soumav Biswas, Rayhan Hayther Samrat, Tafhimul Islam Akash, Arik Subhana, Celia Shahnaz EMBC, 2022 (oral presentation) A CNN-LSTM hybrid architecture is proposed that is capable of mapping PPG signals to corresponding capnograph sequences.
	A Deep Learning Scheme for Detecting Atrial Fibrillation Based on Fusion of Raw and Discrete Wavelet Transformed ECG Features Md Awsafur Rahman, Shahed Ahmed, Shaikh Anowarul Fattah EMBC, 2022 We demonstrate that using the DWT features alongside the original ECG signal helps improve Atrial Fibrillation (AF) classification performance.
	An Approach for Analyzing Cognitive Behavior of Autism Spectrum Disorder Using P300 BCI Data Nabila Tasnim, Joyita Halder, Shahed Ahmed, Shaikh Anowarul Fattah IEEE Region 10 Symposium (Tensymp), 2022 We perform analysis on the EEG data of ASD patients going through rehabilitation from the BCIAUT-P300 dataset and report some interesting trends.
	A Siamese Based One Shot Learning Network with a Watermark Enhancement Technique for Historical Watermark Recognition Sawradip Saha, Utsab Saha, Swojan Datta Sammya, Shahed Ahmed, Shaikh Anowarul Fattah IEEE Region 10 Symposium (Tensymp), 2022 An image pre-processing technique is used to improve the classification accuracy of a Siamese-based one-shot learning framework.
	DSWE-Net: A deep learning approach for shear wave elastography and lesion segmentation using single push acoustic radiation force Shahed Ahmed, Uday Kamal, Md.Kamrul Hasan Ultrasonics, 2021 The proposed DSWE-Net has shown great promise in producing Shear Wave Elastography (SWE) images from scatterer motion data induced by a single Acoustic Radiation Force (ARF) push.

Website adapted from here.