Method for fetal ultrasound image classification using pseudo-labelling with PCA-KMeans and an attention-augmented MobileNet-LSTM model

Accurate classification of fetal ultrasound images is critical for early diagnosis, yet remains challenging due to limited labeled data and high inter-class variability. This study presents a robust deep learning framework that combines a MobileNet backbone with multi-head self-attention and LSTM la...

Full description

Saved in:
Bibliographic Details
Main Authors: Aniket K. Shahade, Priyanka V. Deshmukh, Pritam H. Gohatre, Kanchan S. Tidke, Rohan Ingle
Format: Article
Language:English
Published: Elsevier 2025-12-01
Series:MethodsX
Subjects:
Online Access:http://www.sciencedirect.com/science/article/pii/S2215016125004078
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849233750499524608
author Aniket K. Shahade
Priyanka V. Deshmukh
Pritam H. Gohatre
Kanchan S. Tidke
Rohan Ingle
author_facet Aniket K. Shahade
Priyanka V. Deshmukh
Pritam H. Gohatre
Kanchan S. Tidke
Rohan Ingle
author_sort Aniket K. Shahade
collection DOAJ
description Accurate classification of fetal ultrasound images is critical for early diagnosis, yet remains challenging due to limited labeled data and high inter-class variability. This study presents a robust deep learning framework that combines a MobileNet backbone with multi-head self-attention and LSTM layers to enhance feature learning and temporal context. To address data scarcity and imbalance, unsupervised clustering was employed using Principal Component Analysis (PCA) for dimensionality reduction and K-means (k=4) for pseudo-label generation. These pseudo-labeled clusters were then balanced using oversampling techniques. The proposed model was trained using transfer learning on the augmented dataset and achieved a test accuracy of approximately 98 % with a macro-F1 score of 0.98, indicating highly reliable classification performance. • Employed PCA (100 components) and K-means (k=4) for effective pseudo-labeling and class balancing. • Designed a hybrid deep learning architecture using MobileNet, multi-head attention, and LSTM. • Achieved ∼98 % test accuracy and 0.98 macro-F1 score, demonstrating strong model generalization.
format Article
id doaj-art-c8f8294da73c49c2ba7c4b8f591ba294
institution Kabale University
issn 2215-0161
language English
publishDate 2025-12-01
publisher Elsevier
record_format Article
series MethodsX
spelling doaj-art-c8f8294da73c49c2ba7c4b8f591ba2942025-08-20T04:03:25ZengElsevierMethodsX2215-01612025-12-011510356310.1016/j.mex.2025.103563Method for fetal ultrasound image classification using pseudo-labelling with PCA-KMeans and an attention-augmented MobileNet-LSTM modelAniket K. Shahade0Priyanka V. Deshmukh1Pritam H. Gohatre2Kanchan S. Tidke3Rohan Ingle4Symbiosis Institute of Technology, Pune Campus, Symbiosis International (Deemed University), Pune, India; Corresponding author.Symbiosis Institute of Technology, Pune Campus, Symbiosis International (Deemed University), Pune, IndiaVisvesvaraya National Institute of Technology, Nagpur, IndiaDr. Rajendra Gode Institute of Technology & Research, Amravati, IndiaSymbiosis Institute of Technology, Pune Campus, Symbiosis International (Deemed University), Pune, IndiaAccurate classification of fetal ultrasound images is critical for early diagnosis, yet remains challenging due to limited labeled data and high inter-class variability. This study presents a robust deep learning framework that combines a MobileNet backbone with multi-head self-attention and LSTM layers to enhance feature learning and temporal context. To address data scarcity and imbalance, unsupervised clustering was employed using Principal Component Analysis (PCA) for dimensionality reduction and K-means (k=4) for pseudo-label generation. These pseudo-labeled clusters were then balanced using oversampling techniques. The proposed model was trained using transfer learning on the augmented dataset and achieved a test accuracy of approximately 98 % with a macro-F1 score of 0.98, indicating highly reliable classification performance. • Employed PCA (100 components) and K-means (k=4) for effective pseudo-labeling and class balancing. • Designed a hybrid deep learning architecture using MobileNet, multi-head attention, and LSTM. • Achieved ∼98 % test accuracy and 0.98 macro-F1 score, demonstrating strong model generalization.http://www.sciencedirect.com/science/article/pii/S2215016125004078Fetal ultrasoundImage classificationPseudo-labelingPCAK-means clusteringMobileNet
spellingShingle Aniket K. Shahade
Priyanka V. Deshmukh
Pritam H. Gohatre
Kanchan S. Tidke
Rohan Ingle
Method for fetal ultrasound image classification using pseudo-labelling with PCA-KMeans and an attention-augmented MobileNet-LSTM model
MethodsX
Fetal ultrasound
Image classification
Pseudo-labeling
PCA
K-means clustering
MobileNet
title Method for fetal ultrasound image classification using pseudo-labelling with PCA-KMeans and an attention-augmented MobileNet-LSTM model
title_full Method for fetal ultrasound image classification using pseudo-labelling with PCA-KMeans and an attention-augmented MobileNet-LSTM model
title_fullStr Method for fetal ultrasound image classification using pseudo-labelling with PCA-KMeans and an attention-augmented MobileNet-LSTM model
title_full_unstemmed Method for fetal ultrasound image classification using pseudo-labelling with PCA-KMeans and an attention-augmented MobileNet-LSTM model
title_short Method for fetal ultrasound image classification using pseudo-labelling with PCA-KMeans and an attention-augmented MobileNet-LSTM model
title_sort method for fetal ultrasound image classification using pseudo labelling with pca kmeans and an attention augmented mobilenet lstm model
topic Fetal ultrasound
Image classification
Pseudo-labeling
PCA
K-means clustering
MobileNet
url http://www.sciencedirect.com/science/article/pii/S2215016125004078
work_keys_str_mv AT aniketkshahade methodforfetalultrasoundimageclassificationusingpseudolabellingwithpcakmeansandanattentionaugmentedmobilenetlstmmodel
AT priyankavdeshmukh methodforfetalultrasoundimageclassificationusingpseudolabellingwithpcakmeansandanattentionaugmentedmobilenetlstmmodel
AT pritamhgohatre methodforfetalultrasoundimageclassificationusingpseudolabellingwithpcakmeansandanattentionaugmentedmobilenetlstmmodel
AT kanchanstidke methodforfetalultrasoundimageclassificationusingpseudolabellingwithpcakmeansandanattentionaugmentedmobilenetlstmmodel
AT rohaningle methodforfetalultrasoundimageclassificationusingpseudolabellingwithpcakmeansandanattentionaugmentedmobilenetlstmmodel