Publications
M. B. Bijoy et al., “Deep Cleaner—A Few Shot Image Dataset Cleaner Using Supervised Contrastive Learning,” in IEEE Access, vol. 11, pp. 18727-18738, 2023, doi: 10.1109/ACCESS.2023.3247500.
Data collection is inherently error-prone, often necessitating pre-processing before feeding it into machine learning models. Our paper introduces a pioneering approach to data cleaning through self-supervised learning. What sets this method apart is not only its effectiveness in cleansing datasets but also its ability to train models with few-shot capabilities, enhancing their adaptability.
M. B. Bijoy et al., “Cervix type detection using a self-supervision boosted object detection technique” in International Journal of Imaging systems and technology, vol. 32, issue 5, doi: 10.1002/ima.22696. (Wiley)
Cervical cancer, a significant global health concern, inspired us to create an advanced detection model, EfficientCenterDet, employing self-supervision techniques. This model outperforms traditional methods by achieving a remarkable 10% improvement in accuracy on MobileODT cervical data. Additionally, our approach significantly enhances cervical type classification, reaching an impressive accuracy rate of 87%.
Limna Das, P et al., Early Detection of COVID-19 from CT Scans Using Deep Learning Techniques. In: Thampi, S.M., Gelenbe, E., Atiquzzaman, M., Chaudhary, V., Li, KC. (eds) Advances in Computing and Network Communications. Lecture Notes in Electrical Engineering, vol 736. Springer, Singapore. (Springer)
The global COVID-19 pandemic has exposed the need for quick and reliable diagnostic tools. Our solution leverages deep learning algorithms, particularly the EfficientNet architecture, to detect COVID-19 from CT scans with a 10% accuracy boost compared to ResNet. We're on a mission to deploy this as a chatbot, continuously improving its accuracy through ongoing learning.