Deep Learning-Based Lung Medical Image Recognition
Journal Title: International Journal of Innovative Research in Computer Science and Technology - Year 2024, Vol 12, Issue 3
Abstract
Pulmonary nodules serve as critical indicators for early lung cancer diagnosis, making their detection and classification essential. The prevalent use of transfer learning in recognition algorithms often encounters a significant disparity between source and target datasets, which hampers effective feature extraction from pulmonary nodules and degrades performance. An enhanced neural network model leveraging convolutional neural networks is introduced to address this issue. This model integrates a pre-trained GoogLeNet Inception V3 network with a custom-designed feature fusion layer, improving the network’s ability to extract features. To ascertain the optimal configuration, the models were evaluated based on accuracy in various combinations. The experiments conducted on the LUNA16 pulmonary nodule dataset revealed that the refined network model achieved an accuracy of 88.78% and a sensitivity of 87.18%. This represents an increase of 2.7 and 2.22 percentage points in accuracy and sensitivity, respectively, compared to the GoogLeNet Inception V3 algorithm. Further tests across different dataset proportions also yielded superior outcomes, demonstrating enhanced generalization capabilities. These findings can offer objective benchmarks for clinical diagnosis.
Authors and Affiliations
Xinghui Fei Yulu Wang Lu Dai Mingxiu Sui
Keywords
Related Articles
- EP ID EP744979
- DOI 10.55524/ijircst.2024.12.3.16
- Views 68
- Downloads 0
A Review Paper on Digital Marketing
The promotion of goods or brands through the internet is known as digital marketing. In current years, the growth in electronic advertising in guesthouses has increased, with illustrations including Facebook Marketing th...
Design And Development of Sketch Based Image Retrieval Using Deep Learning
In this cutting edge, the common wrong doing rate is expanding day-by-day and to manage up with this the criminal divisions as well ought to discover ways in which would speed up the by and large preparation and offer as...
Performance Enhancement of an Integrated Offshore Wind Farm, Marine Current Farm and PV System with Flywheel Energy Storage System
This paper represents a control strategy to instantaneously realize mitigation in power fluctuation and performance enhancement of offshore wind farm (OWF), PV system and Marine current farm (MCF) coupled to the grid wit...
Optimization Techniques for Fuzzy-PID Control of CSC-Based D-STATCOM for Power Quality Improvement
This study focuses on enhancing power quality using a Current Source Converter (CSC) based Dynamic Voltage Restorer (D-STATCOM) controlled by a Fuzzy Logic-PID (Fuzzy-PID) controller. Power quality improvement is a vital...
Comparative Analysis of Linear and Digitalized Thermal Wave Imaging Methods
In order to identify surface features of an object, active thermal non-destructive testing (TNDT) techniques use the surface temperature profile that has been measured over the object. To improve the flaw detectability u...