A Multi-Task Attention-Driven SegNet for Lung Infection Segmentation and Classification From HRCT Images

Also Available Domains Deep Learning

Project Code :TCMAPY2505

Objective

This project builds a multi-task deep learning framework using MS-CSTA-Net and DualMask-SegNet v5.4 to jointly segment and classify lung infections from HRCT images via multi-scale attention and dual mask learning. A web interface allows image upload, segmentation map viewing, and classification results, with performance evaluated using accuracy, precision, recall, F1-score, and IoU. The scalable system adapts to diverse datasets, reduces manual effort, and offers interpretable infection visualizations to support medical imaging research.

Abstract

Digital dependency is characterized by excessive engagement with digital devices, leading to psychological and behavioral changes. This project develops a machine learning based system for psychological profiling of digital dependency using multiple regression algorithms. The system collects user inputs across behavioral and demographic attributes and applies trained models (Linear Regression, Ridge, Lasso, Random Forest, XGBoost, LightGBM, DeepGBM, DCN‑V2, and Mixture of Experts) to predict a dependency score on a scale of 0–100. Separate modules are implemented for emotion detection and posture analysis. A Flask web application provides user registration, login, model selection, prediction input, result display, and logout. The system enables researchers and individuals to quantify digital dependency and access supportive feedback.

Keywords: Digital dependency, psychological profiling, machine learning, DeepGBM, DCN‑V2, Mixture of Experts, XGBoost, Flask, regression, web application.

NOTE: Without the concern of our team, please don't submit to the college. This Abstract varies based on student requirements.