MUMBAI, India, Aug. 17 -- Intellectual Property India has published a patent application (202641095061 A) filed by Adhiparasakthi Engineering College, Melmaruvathur on August 05, 2026, for Leveraging Ai And Deep Learning For Improved Psoriasis Variant Recognition.
Inventors include Dr. C. Dhaya; Mr. K. Chairmadurai; A. Thamarai Selvi; Dr. R. Srivel; Mr. G. Srinivasan; and Mr. S. Hariganesan.
The application for the patent was published on August 14, 2026, under issue no. 33/2026.
Abstract: Psoriasis is a chronic autoimmune skin disorder that manifests in five clinically distinct variants — plaque, guttate, inverse, pustular, and erythrodermic — each requiring specific treatment strategies. Accurate automated classification of psoriasis variants from dermatological images remains challenging due to inter-class visual similarity and intra-class variability across patients. This invention presents a deep learning-based psoriasis variant recognition system employing a dual-branch hybrid architecture that combines a custom Convolutional Neural Network (CNN) and a ResNet-50 transfer learning model. Input skin images are preprocessed through resizing (224×224 pixels), normalization, noise reduction, and data augmentation (rotation, flipping, zooming, brightness adjustment). The CNN branch extracts hierarchical local features through three Conv2D + ReLU + MaxPooling + BatchNormalization blocks, while the ResNet-50 branch (pretrained on ImageNet with selective last-30-layer fine-tuning) provides deep se
Disclaimer: Curated by HT Syndication.
