Investigates the critical challenge of domain shift by framing a deep learning model as a scientific instrument — quantifying why and how pathology models fail on unseen clinical data. Also includes a tool to visualise segmented histology images and phenotype data.
Investigates whether unsupervised clustering can automatically group continuous 3D human movement coordinates into discrete, identifiable poses, automating the data annotation pipeline. Uses Indian Classical Dance (Bharatanatyam) as a controlled ground-truth dataset. Features a custom dual-stage heuristic filter to eliminate "Pose Hallucination" artifacts from MediaPipe.
Comparative evaluation of three distinct unsupervised clustering algorithms (K-Means, GMM, FCM) at superpixel, pixel, and stain-specific levels for nuclei segmentation in H&E-stained images from the MoNuSeg dataset. Offers a scalable alternative to supervised techniques.
Systematic evaluation of four GAN architectures (Standard GAN, DCGAN, DualGAN, WGAN-GP) to synthesize realistic EEG signals for two mental states. WGAN-GP achieved top combined performance (0.6623), validating GANs as a data augmentation strategy for neurophysiological research.
M. N. Reddy E and L. B. Rananavare, "Deep Learning Approaches for Plant Disease Management: A State-of-the-Art Analysis," 2023 7th International Conference on Computation System and Information Technology for Sustainable Solutions (CSITSS), Bangalore, India, 2023, pp. 1–5, doi: 10.1109/CSITSS60515.2023.10334120.
M. N. Reddy, V.V. Badami, Dinesh B, S. P. Chetan Kumar, and K.V. Kanzaria, "Spoken English to Indian Sign Language Translator," in Proceedings of the 4th International Virtual Conference on Advances in Computing & Information Technology (IACIT-2022), River Publishers, 2022, pp. 29–39.
Proceedings → PDF →M. N. Reddy E, P. Moggridge, "A Systematic Survey and Taxonomy of Deep Learning for Nuclei Segmentation in H&E Histopathology."
M. N. Reddy E, P. Moggridge, et al., "An Empirical Investigation into the Domain Generalization of Nuclei Segmentation and Phenotyping Models in Computational Histopathology."