Intellectual Property India Publishes Patent Application for Hybrid Cnn-Based Deep Learning System For Quality Detection Of Submersible Pump Impeller Using Resnet50 And Mobilenetv2

MUMBAI, India, June 22 -- Intellectual Property India has published a patent application (202641069332 A) filed by Easwari Engineering College on June 03, 2026, for Hybrid Cnn-Based Deep Learning System For Quality Detection Of Submersible Pump Impeller Using Resnet50 And Mobilenetv2.

Inventors include Uma Maheshwari S; Sai Sachin R; and Rishikesh R.

The application for the patent was published on June 12, 2026, under issue no. 24/2026.

Abstract: Ensuring product quality during export transportation is essential to maintain trade standards and reduce economic losses. This invention proposes an automated inspection framework for casting products using a hybrid deep learning model. The system analyzes images of defective and high-quality castings from an industrial dataset. A combined ResNet50 and MobileNetV2 architecture extracts complementary features to improve defect detection accuracy and robustness. The model classifies products for export readiness and is deployed through a Streamlit web application for real-time inspection. The approach enhances speed, consistency, and reliability in industrial quality assurance. Total Number of Words in Abstract:90

Disclaimer: Curated by HT Syndication.