Interests

Medical Image Analysis

Efficient Deep Learning

Multimodal AI

Computer Vision for Social Good

My research focuses on addressing the following key questions:

1. How can we create a storage-efficient solution to capture both local and global features from medical images?

2. How can we efficiently capture relationships between different modalities in low-resourced settings?

3. How can we efficiently generate images in sustainable medical imaging solutions?

4. How can we efficiently build Computer Vision solutions in resource-constrained and federated frameworks?

5. How do we algorithmically optimize Optical Neural Networks to make them more efficient?

Publications and Research Works

Research Featured in:

CVAMD@ICCV2025

WVLL@WACV2024

Selected Research Highlights:

BAMPolyp: Bi-Axial Mamba Bottleneck for Gastrointestinal Polyp Segmentation

Bi-Axial Mamba module applies state-space mixing along height and width axes independently. This design enables each spatial token to capture global dependencies along one axis while preserving local spatial continuity along the other, effectively enhancing both boundary sensitivity and contextual coherence, critical for distinguishing polyps from mucosal folds. An EfficientNet-B0 encoder and deep supervision decoder ensure strong hierarchical feature extraction and optimization. Ablation studies confirm that axis-wise decomposition improves localization accuracy without compromising semantic understanding. Our results establish BAMPolyp as a robust and efficient solution for GI polyp segmentation, achieving 0.9380/0.8888 (Dice/IoU) on Kvasir-SEG, 0.9437/0.8939 on CVC-ClinicDB, 0.9255/0.8659 on CVC-ColonDB, and 0.8683/0.8211 on PolypGen with minimal computational overhead.

Accepted: CVAMD@ICCV2025!!
(MICCAI endorsed CVAMD, ICCV: A*)

Involution-based efficient autoencoder for denoising histopathological images with enhanced hybrid feature extraction

We propose a lightweight autoencoder (43.11 kilobytes) for denoising histopathology images by fusing a single involution layer within a small convolution model, resulting in better denoising performance in a hybrid model, which has both channel-specific and location-specific feature extraction capabilities. Building upon the idea of a shallow autoencoder, our model results in much lower memory and compute overhead requirements, while also not avoiding the generation of artifacts. On Malaria Blood Smear and CRC datasets, SSIM Loss and Peak-Signal-to-Noise-Ratio were used for performance evaluation, with lower SSIM Loss (0.058 and 0.34) in denoising images with an added Gaussian noise of 0.3. Our proposed autoencoder, with low weight parameters of 11,037 and 81,630,000 floating point operations (FLOPs), is over 20 times less computationally expensive than Xception, the second-best performing model, establishing ours as the most efficient denoising autoencoder for histopathology images.

Published: Computers in Biology and Medicine!!
(Q1, Impact Factor: 7.0)

Involution fused convolution for classifying eye-tracking patterns of children with Autism Spectrum Disorder

We utilize two image-processing operations to see how these processes learn eye-tracking patterns. Since these patterns are primarily based on spatial information, we employ a hybrid of involution and convolution. Our study shows that adding a few involution layers reduces size and computational cost while enhancing location-specific capabilities, maintaining performance comparable to pure convolutional models. However, excessive involution layers lead to weaker performance. For comparisons, we experiment with two datasets and a combined version of both in ablation studies. Our proposed model, featuring three involution layers and three convolution layers, achieved 99.43% accuracy on the first dataset and 96.78% on the second, with a size of only 1.36 megabytes. These results showcase the effectiveness of combining involution and convolution layers which outperforms previous literature.

Published: Engineering Applications of Artificial Intelligence!!
(Q1, Impact Factor: 7.5)

Uncovering local aggregated air quality index with smartphone captured images leveraging efficient deep convolutional neural network

Our research involves the development of a Deep Convolutional Neural Network (DCNN), which we train using over a thousand outdoor images taken and annotated. These photos are captured at various locations in Dhaka, and their labels are based on PM2.5 concentration data obtained from the local US consulate, calculated using the NowCast algorithm. Through supervised learning, our model establishes a correlation index during training, enhancing its ability to function as a Picture-based Predictor of PM2.5 Concentration (PPPC). This enables the algorithm to calculate an equivalent daily averaged AQI index from a smartphone image. Unlike, popular overly parameterized models, our model shows resource efficiency since it uses fewer parameters. Furthermore, test results indicate that our model outperforms popular models like ViT and INN, as well as popular CNN-based models such as VGG19, ResNet50, and MobileNetV2, in predicting location-specific PM2.5 concentration. Our dataset is the first publicly available collection that includes atmospheric images and corresponding PM2.5 measurements from Dhaka. Our codes and dataset are available at https://github.com/lepotatoguy/aqi.

Published: Scientific Reports!!
🎖️ Sustainability Top 100 of 2024!!
(Q1, Impact Factor: 3.8)

List of Publication:

Check updates in Google Scholar: https://scholar.google.com/citations?user=4a4covIAAAAJ&hl

ORCID: 0000-0003-3249-4490

Journal Articles:

Islam, Md. F., Reza, Md. T., Manab, M. A., Zabeen, S., Islam, Md. F. U., Shahriar, Md. F., Kaykobad, M., & Noor, J. (2025). Involution-based efficient autoencoder for denoising histopathological images with enhanced hybrid feature extraction. Computers in Biology and Medicine, 192, 110174. https://doi.org/10.1016/j.compbiomed.2025.110174 [Q1, IF: 7.0]
Islam, Md. F., Manab, M. A., Mondal, J. J., Zabeen, S., Rahman, F. B., Hasan, Md. Z., Sadeque, F., & Noor, J. (2025). Involution fused convolution for classifying eye-tracking patterns of children with Autism Spectrum Disorder. Engineering Applications of Artificial Intelligence, 139, 109475. https://doi.org/10.1016/j.engappai.2024.109475 [Q1, IF: 7.5]
Mondal, J. J., Islam, Md. F., Islam, R., Rhidi, N. K., Newaz, S., Islam, A. B. M. A. A., Manab, M. A., & Noor, J. (2024). Uncovering local aggregated air quality index with smartphone captured images leveraging efficient deep convolutional neural network. Scientific Reports. https://doi.org/10.1038/s41598-023-51015-1. [Q1, IF: 3.8]

Conference or Workshop Papers:

Islam, Md. F., Ahmed, T., Chanda, P., Mondal, J. J., Manab, M. A., Zabeen, S. & Noor, J. (2025). BAMPolyp: Bi-Axial Mamba Bottleneck for Gastrointestinal Polyp Segmentation, [Accepted for Publication at CVAMD-ICCV2025 (ERA A, CORE A*)]
Nasim, H., Printia, F. J., Himel, M. H., Rashid, R., Chowdhury, I. J., Mondal, J. J., Islam, Md. F & Noor, J. (2024). Fog-Resilient Bangla Car Plate Recognition using Dark Channel Prior and YOLO, Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision (WACV) Workshops. [ERA A, CORE A]
Zabeen, S., Hasan A., Islam, Md. F, Sabbir Hossain, M., & Rasel, A. A. (2023). Robust Fake Review Detection Using Uncertainty-Aware LSTM and BERT, 2023 15th International Conference on Computational Intelligence and Communication Networks (CICN). 786-791, https://doi.org/10.1109/CICN59264.2023.10402342.
Chy, Md. S. R., Mahin, M. R. H., Islam, Md. F., Hossain, M. S., & Rasel, A. A. (2023). Classifying Corn Leaf Diseases Using Ensemble Learning with Dropout and Stochastic Depth Based Convolutional Networks. Proceedings of the 2023 8th International Conference on Machine Learning Technologies, 185–189. https://doi.org/10.1145/3589883.3589911
Anwar, Md. T., Islam, Md. F., & Alam, Md. G. R. (2023). Forecasting Meteorological Solar Irradiation Using Machine Learning and N-BEATS Architecture. Proceedings of the 2023 8th International Conference on Machine Learning Technologies, 46–53. https://doi.org/10.1145/3589883.3589890
Islam, Md. F., Zabeen, S., Islam, Md. A., Rahman, F. B., Ahmed, A., Karim, D. Z., Rasel, A. A., & Manab, M. A. (2023). How certain are transformers in image classification: uncertainty analysis with Monte Carlo dropout. In W. Osten, D. P. Nikolaev, & J. (Jessica) Zhou (Eds.), Fifteenth International Conference on Machine Vision (ICMV 2022) (Vol. 12701, p. 127010K). SPIE. https://doi.org/10.1117/12.2679442
Islam, Md. F., Zabeen, S., Rahman, F. B., Islam, Md. A., Kibria, F. B., Manab, M. A., Karim, D. Z., & Rasel, A. A. (2023). Exploring Node Classification Uncertainty in Graph Neural Networks. Proceedings of the 2023 ACM Southeast Conference, 186–190. https://doi.org/10.1145/3564746.3587019
Islam, Md. F., Zabeen, S., Rahman, F. B., Islam, Md. A., Kibria, F. B., Manab, M. A., Karim, D. Z., & Rasel, A. A. (2023). UnIC-Net: Uncertainty Aware Involution-Convolution Hybrid Network for Two-level Disease Identification. SoutheastCon 2023, 305–312. https://doi.org/10.1109/SoutheastCon51012.2023.10115109
Rhythm, E. R., Shuvo, R. A., Hossain, M. S., Islam, Md. F., & Rasel, A. A. (2023). Sentiment Analysis of Restaurant Reviews from Bangladeshi Food Delivery Apps. 2023 International Conference on Emerging Smart Computing and Informatics (ESCI), 1–5. https://doi.org/10.1109/ESCI56872.2023.10100214
Mumenin, N., Islam, Md. F., Chowdhury, Md. R. Z., & Yousuf, M. A. (2023). Diagnosis of Autism Spectrum Disorder Through Eye Movement Tracking Using Deep Learning. In M. Ahmad, M. S. Uddin, & Y. M. Jang (Eds.), Proceedings of International Conference on Information and Communication Technology for Development (pp. 251–262). Springer Nature Singapore.
Islam, Md. F., Bin Rahman, F., Zabeen, S., Islam, Md. A., Sabbir Hossain, M., Kabir Mehedi, M. H., Manab, M. A., & Rasel, A. A. (2022). RNN Variants vs Transformer Variants: Uncertainty in Text Classification with Monte Carlo Dropout. 2022 25th International Conference on Computer and Information Technology (ICCIT), 7–12. https://doi.org/10.1109/ICCIT57492.2022.10055922
Islam, Md. F., Zabeen, S., Rahman, M. M., Khan, M. H., Khan, F. N., Nahim, N. Z., Anwar, T., & Kaykobad, M. (2022). Identifying Hurricane Damage using Explainable Compact Transformer with Convolutional Embedding. 2022 25th International Conference on Computer and Information Technology (ICCIT), 833–838. https://doi.org/10.1109/ICCIT57492.2022.10054917
Mondal, J. J., Islam, Md. F., Zabeen, S., & Manab, M. A. (2022). InvoPotNet: Detecting Pothole from Images through Leveraging Lightweight Involutional Neural Network. 2022 25th International Conference on Computer and Information Technology (ICCIT), 599–604. https://doi.org/10.1109/ICCIT57492.2022.10055818
Islam, Md. F., Zabeen, S., Mehedi, Md. H. K., Iqbal, S., & Rasel, A. A. (2022). Monte Carlo Dropout for Uncertainty Analysis and ECG Trace Image Classification. In A. Krzyzak, C. Y. Suen, A. Torsello, & N. Nobile (Eds.), Structural, Syntactic, and Statistical Pattern Recognition (pp. 173–182). Springer International Publishing. [ERA A, CORE B]
Mondal, J. J., Islam, Md. F., Zabeen, S., Islam, A. B. M. A. A., & Noor, J. (2022). Note: Plant Leaf Disease Network (PLeaD-Net): Identifying Plant Leaf Diseases through Leveraging Limited-Resource Deep Convolutional Neural Network. Proceedings of the 5th ACM SIGCAS/SIGCHI Conference on Computing and Sustainable Societies, 668–673. https://doi.org/10.1145/3530190.3534844

Preprints:

Islam, Md. F., Reza, Md. T., Manab, M. A., Mahin, M. R. H., Zabeen, S., & Noor, J. (2024). Spatially Optimized Compact Deep Metric Learning Model for Similarity Search. arXiv preprint. https://doi.org/10.48550/arXiv.2404.06593.
Islam, Md. F., Zabeen, S., Manab, M.A., Mahin, M.R.H., Mondal, J.J., Reza, M.T., Hasan, M.Z., Haque, M., Sadeque, F. and Noor, J. (2024). Med-IC: Fusing a Single Layer Involution with Convolutions for Enhanced Medical Image Classification and Segmentation. arXiv preprint arXiv:2409.18506.
Parsa, A. F., Abdullah, S. M., Talukdar, A. H., Kabbya, Md. A. S., Hasan, S. A., Islam, Md. F. and Noor, J. “A Comparative Performance Analysis of Classification and Segmentation Models on Bangladeshi Pothole Dataset”, arXiv preprint, DOI: 10.48550/arXiv:2501.06602.
T. Ahmed, S. Jannat, M. F. Islam, & J. Noor, “Involution-Infused DenseNet with Two-Step Compression for Resource-Efficient Plant Disease Classification”, arXiv preprint, DOI: 10.48550/arXiv.2506.00735.

Undergraduate Thesis:

Islam, Md. F., Zabeen, S., Rahman, F. B., Islam, Md. A., Kibria, F. B. (2023). Analysis of Uncertainty in Different Neural Network Structures using Monte Carlo Dropout. BRAC University. https://dspace.bracu.ac.bd/xmlui/handle/10361/22153 . (Supervisor: Annajiat Alim Rasel, Co-supervisors: Dewan Ziaul Karim & Meem Arafat Manab.)

Research Posters:

Nasim, H., Printia, F. J., Himel, M. H., Rashid, R., Chowdhury, I. J., Islam, Md. F & Noor, J. (2023). Fog-Resilient Bangla Car Plate Recognition using Dark Channel Prior and YOLO. 10th International Conference on Networking, Systems and Security. https://doi.org/10.13140/RG.2.2.10681.13925.
Mondal, J. J., Islam, Md. F., Islam, R., Rhidi, N. K., Manab, M. A., Islam, A. B. M. A. A & Noor, J. (2022). Unmasking the Invisible: Finding Location-Specific Aggregated Air Quality Index with Smartphone Images. 9th International Conference on Networking, Systems and Security. https://doi.org/10.13140/RG.2.2.12552.08968.
Mondal, J. J., Islam, Md. F., Zabeen, S., Islam, A. B. M. A. A., & Noor, J. (2021). Identification of Plant Leaf Diseases using Deep Convolutional Neural Network with Less Computational Power. 8th International Conference on Networking, Systems and Security. https://doi.org/10.13140/RG.2.2.17702.04168.

Funded Projects

Data Ghurhi

A Bengali-based Quantitative Data Collection and Analysis Tool by RIC, BUET

Funded by: The World Bank, ICT Division, Bangladesh.

My Contributions:

Developed Python codes for statistical tests, visualization, and data processing of the tool.
Developed an automated Bengali label-based graph plotting program that properly displays the Bengali words.
Easy to do statistical tests with instructions and automated data processing with multiple options for the user.
User-centric customizable graph visualization in both English and Bengali.
Machine Learning prediction tests (under development)
Logo designing and conceptualization.

Peer Review and Other Services

Peer Reviewing:

ACM Journal on Computing and Sustainable Societies (ACM JCSS)
ICCV 2023 Workshop on Resource Efficient Deep Learning for Computer Vision (RCV 2023)
WACV 2024 Workshop on Vision-Based Understanding for Low-Resource Languages (WVLL 2024)
IEEE International Joint Conference on Neural Networks (IJCNN 2024)
The 27th ACM Conference on Computer-Supported Cooperative Work and Social Computing (CSCW 2024)
Asia-Pacific Conference on Computer Science and Data Engineering 2023 (CSDE 2023)
International Conference on Networks, Communication and Information Technology (NCIT 2022)
International Journal of Computing and Digital Systems (IJCDS)
Nature Scientific Reports
The Journal of Supercomputing
Program Committee: The 40th Annual AAAI Conference on Artificial Intelligence (AAAI 2026)
Discover Applied Sciences

BRACU Research Service:

Pre-Thesis 1 Panel Member/Judge
Pre-Thesis 2 Panel Member/Judge

Page updated

Report abuse