MUMBAI, India, Aug. 12 -- Intellectual Property India has published a patent application (202641092985 A) filed by Dr. R. Dhanalakshmi; and Gokulakrishnan. S on July 31, 2026, for System And Method For Stain-Invariant Feature Extraction In Histopathological Images Using Adversarial Domain-Agnostic Representation Learning.
Inventors include Dr. R. Dhanalakshmi; and Gokulakrishnan. S.
The application for the patent was published on August 07, 2026, under issue no. 32/2026.
Abstract: ABSTRACT A system and method for stain-invariant feature extraction from histopathological images is disclosed. The system comprises a feature extractor network (103) operatively coupled, through a gradient reversal layer (105), to a stain discriminator network (104), such that the feature extractor network (103) is adversarially trained to generate feature representations (106) from which the originating staining protocol or laboratory-of-origin cannot be reliably predicted. The resulting Stain- Invariant Feature Extractor (SIFE) module (102) may be deployed as a frozen, plug-in front-end module to any downstream diagnostic classifier (107), enabling generalization across previously unseen laboratories and staining protocols without retraining, explicit colour normalization, or paired stain-transfer data, thereby addressing a critical bottleneck in the scalable clinical deployment of AI-assisted histopathological diagnosis. (Reference Figure for Abstract: FIG. 1)
Disclaimer: Curated by HT Syndication.
