An Improved Deep Learning Models with Hybrid Architectures Thyroid Disease Classification Diagnosis
Year: 2025
Type: article
Abstract: Diagnosing thyroid disease is challenging because the disease presents itself through a spectrum of subtle and diverse symptoms. The study presents a refined deep learning method that utilizes a combination of Convolutional Neural Networks (CNNs) and Recurrent Neural Networks (RNNs) to improve both diagnostic accuracy and efficiency. The CNN module extracts spatial features from thyroid ultrasound images, and the RNN module analyzes these features in sequences to identify temporal patterns that can reveal the progression or type of thyroid conditions. The performance of our model against a dataset of labelled ultrasound images and patient data demonstrates notable advancements in classification and diagnostic accuracy compared to conventional techniques. Combining CNN and RNN architectures delivers a powerful approach that enables the automatic detection and classification of thyroid diseases, which leads to enhanced reliability and speed in healthcare diagnostics.
Source: Journal of Neonatal Surgery
Authors: Minal Chaphekar, Omprakash Chandrakar
Language: English
Topic: AI in cancer detection
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: diamond
OpenAlex ID: W4408208002