Deep Learning-based Feature Extraction Research Articles

Enhancing ECG signal classification through pre-trained stacked-CNN embeddings: a transfer learning approach

Rapid and accurate electrocardiogram (ECG) signal classification is crucial in high-stakes healthcare settings. However, existing computational models often struggle to balance high performance with computational efficiency. This study introduces an innovative computational framework that combines transfer learning with traditional machine learning to optimize ECG classification. We use a pre-trained Stacked Convolutional Neural Network (SCNN) to generate high-dimensional feature embeddings, which are then evaluated by an array of machine learning classifiers. Our models demonstrate exceptional performance, particularly when utilizing embeddings from SCNNs trained on diverse datasets. This underscores the importance of data diversity in improving classifier discrimination. Notably, Multilayer Perceptrons (MLPs) stand out for their ability to balance computational efficiency with strong performance, achieving test F1-scores of 0.94 and 1.00 in multi-class and binary tasks on the CinC2017 dataset, and 0.85 and 0.99 on the CPSC2018 dataset. Our approach consistently outperforms existing methods, setting new benchmarks in ECG classification. The synergy between deep learning-based feature extraction and traditional machine learning through transfer learning offers a robust, efficient, and adaptable strategy for ECG classification, addressing a critical research gap and laying the groundwork for future advancements in this crucial healthcare field.

Chilly Leaves Diseases Identification Using MobileNet V2 Deep Learning Model

This work describes the application of pre-trained deep learning models (Mobile Net V1, Mobile Net V2) to image classification issues such as plant disease diagnosis. Regarding this, We provide a deep learning-based feature extraction technique for classifying leaf diseases of cold plants and identifying plant species. The automated diagnosis of plant diseases has brought about a transformation in the agriculture industry. Our main goal is to use neural networks for agricultural disease diagnosis and treatment. Plant diseases seriously impair the agriculture industry's financial standing. Managing an illness is a challenging endeavor. Usually, infections appear on the leaves or stems of plants, as do their symptoms, which include colorful spots or steaks. Compared to human labor, image processing technology enables faster and more accurate illness detection. Image processing is critical in the detection of plant diseases because it gives the best results with the least amount of human intervention. Deep learning helps to diagnose diseases and provide focused treatment.

Two-Level Spatio-Temporal Feature Fused Two-Stream Network for Micro-Expression Recognition.

Micro-expressions, which are spontaneous and difficult to suppress, reveal a person's true emotions. They are characterized by short duration and low intensity, making the task of micro-expression recognition challenging in the field of emotion computing. In recent years, deep learning-based feature extraction and fusion techniques have been widely used for micro-expression recognition, particularly methods based on Vision Transformer that have gained popularity. However, the Vision Transformer-based architecture used in micro-expression recognition involves a significant amount of invalid computation. Additionally, in the traditional two-stream architecture, although separate streams are combined through late fusion, only the output features from the deepest level of the network are utilized for classification, thus limiting the network's ability to capture subtle details due to the lack of fine-grained information. To address these issues, we propose a new two-level spatio-temporal feature fused with a two-stream architecture. This architecture includes a spatial encoder (modified ResNet) for learning texture features of the face, a temporal encoder (Swin Transformer) for learning facial muscle motor features, a feature fusion algorithm for integrating multi-level spatio-temporal features, a classification head, and a weighted average operator for temporal aggregation. The two-stream architecture has the advantage of extracting richer features compared to the single-stream architecture, leading to improved performance. The shifted window scheme of Swin Transformer restricts self-attention computation to non-overlapping local windows and allows cross-window connections, significantly improving the performance and reducing the computation compared to Vision Transformer. Moreover, the modified ResNet is computationally less intensive. Our proposed feature fusion algorithm leverages the similarity in output feature shapes at each stage of the two streams, enabling the effective fusion of multi-level spatio-temporal features. This algorithm results in an improvement of approximately 4% in both the F1 score and the UAR. Comprehensive evaluations conducted on three widely used spontaneous micro-expression datasets (SMIC-HS, CASME II, and SAMM) consistently demonstrate the superiority of our approach over comparative methods. Notably, our approach achieves a UAR exceeding 0.905 on CASME II, making it one of the few frameworks in the published micro-expression recognition literature to achieve such high performance.

Therapy-induced modulation of tumor vasculature and oxygenation in a murine glioblastoma model quantified by deep learning-based feature extraction

Glioblastoma presents characteristically with an exuberant, poorly functional vasculature that causes malperfusion, hypoxia and necrosis. Despite limited clinical efficacy, anti-angiogenesis resulting in vascular normalization remains a promising therapeutic approach. Yet, fundamental questions concerning anti-angiogenic therapy remain unanswered, partly due to the scale and resolution gap between microscopy and clinical imaging and a lack of quantitative data readouts. To what extend does treatment lead to vessel regression or vessel normalization and does it ameliorate or aggravate hypoxia? Clearly, a better understanding of the underlying mechanisms would greatly benefit the development of desperately needed improved treatment regimens. Here, using orthotopic transplantation of Gli36 cells, a widely used murine glioma model, we present a mesoscopic approach based on light sheet fluorescence microscopic imaging of wholemount stained tumors. Deep learning-based segmentation followed by automated feature extraction allowed quantitative analyses of the entire tumor vasculature and oxygenation statuses. Unexpectedly in this model, the response to both cytotoxic and anti-angiogenic therapy was dominated by vessel normalization with little evidence for vessel regression. Equally surprising, only cytotoxic therapy resulted in a significant alleviation of hypoxia. Taken together, we provide and evaluate a quantitative workflow that addresses some of the most urgent mechanistic questions in anti-angiogenic therapy.

Enhancing breast cancer detection from histopathology images: A novel ensemble approach with deep learning-based feature extraction

Effective detection and diagnostic procedures are necessary to enhance patient results for the common and life-threatening illness of breast cancer. Current approaches have limits in scalability and efficiency, highlighting the need for more study. This work introduces a hybrid Breast Cancer (BC) detecting approach that merges Deep Learning (DL) with pre-trained modeling of Histopathology Images (HPI) and an ensemble-based Machine Learning (ML) approach. DL integration allows learning and identifying hidden trends in intricate BC pictures, while ML techniques provide interpretability and generalization skills. Contrast Limited Adaptive Histogram Equalization (CLAHE) was used on HPI as a pre-processing technique to improve picture quality. The ResNet50V2 model was used for deep feature extraction. The Ensemble Learning (EL) model combines predictions from four basic ML approaches using soft voting. The research attained a superior accuracy, precision, recall, and F1 score compared to the most advanced models. This study provides substantial advancements in breast cancer diagnosis, thorough performance evaluation, and reliable assessment. Furthermore, it helps medical personnel make well-informed choices, enhance patient care, and improve results for BC sufferers.

Deep Reinforcement Learning-Based Feature Extraction and Encoding for Finger-Vein Verification

Deep Reinforcement Learning-Based Feature Extraction and Encoding for Finger-Vein Verification

A new approach based on a 1D + 2D convolutional neural network and evolving fuzzy system for the diagnosis of cardiovascular disease from heart sound signals

Diagnostic methods for cardiovascular disease diagnosis based on heart sound classification have been widely investigated for their noninvasiveness, low-cost, and high efficiency. Most current researches either use manually designed functions or deep learning-based methods to extract features from heart sound signals, but the heart sound signals have highly nonstationary and complex data patterns due to environmental noise and the differences between different stethoscopes. Therefore, using a single feature extraction method does not result in a good feature representation. Moreover, deep learning-based feature extraction methods for heart sound signals usually only use 1D convolution or 2D convolution, which limits the capability of neural networks to extract discriminative features. In addition, many studies do not consider the redundancy of features and the interpretability of decisions, which affects the performance and efficiency of the models. To solve the above problems, this paper first proposes a new convolutional neural network named the 1D + 2D convolutional neural network (1D + 2D-CNN) as a deep learning feature extractor, which combines 1D convolution and 2D convolution. The 1D + 2D-CNN contains two branches, and the feature maps obtained from the two branches are concatenated according to the channel. Then, a 10-layer convolutional network with an attention mechanism is introduced to enhance the feature extraction capability of the network. Second, the advantages and disadvantages when combining deep learning features with manual features in different scenarios are explored. In addition, the mean and variance of each dimensional feature of the dataset are calculated by class and feature selection is achieved by evaluating the importance of each dimension of the feature through a simple statistical formula. Finally, an evolving fuzzy system is used to classify the heart sound signals, as it can provide interpretability for decision-making. For the experimental part, the 2016 PhysioNet/CinC Challenge dataset (PCCD) and our collected publicly available pediatric heart sound dataset (PHSD) are used to evaluate the performance of the model by 10-fold cross-validation. The model in this paper achieves accuracies of 96.3 % and 99.1 % on these two datasets respectively, demonstrating its capability to reach the state-of-the-art level. We also open the code of our algorithms. Diagnostic methods for cardiovascular disease diagnosis based on heart sound classification have been widely investigated for their noninvasiveness, low-cost, and high efficiency. Most current researches either use manually designed functions or deep learning-based methods to extract features from heart sound signals, but the heart sound signals have highly nonstationary and complex data patterns due to environmental noise and the differences between different stethoscopes. Therefore, using a single feature extraction method does not result in a good feature representation. Moreover, deep learning-based feature extraction methods for heart sound signals usually only use 1D convolution or 2D convolution, which limits the capability of neural networks to extract discriminative features. In addition, many studies do not consider the redundancy of features and the interpretability of decisions, which affects the performance and efficiency of the models. To solve the above problems, this paper first proposes a new convolutional neural network named the 1D + 2D convolutional neural network (1D + 2D-CNN) as a deep learning feature extractor, which combines 1D convolution and 2D convolution. The 1D + 2D-CNN contains two branches, and the feature maps obtained from the two branches are concatenated according to the channel. Then, a 10-layer convolutional network with an attention mechanism is introduced to enhance the feature extraction capability of the network. Second, the advantages and disadvantages when combining deep learning features with manual features in different scenarios are explored. In addition, the mean and variance of each dimensional feature of the dataset are calculated by class and feature selection is achieved by evaluating the importance of each dimension of the feature through a simple statistical formula. Finally, an evolving fuzzy system is used to classify the heart sound signals, as it can provide interpretability for decision-making. For the experimental part, the 2016 PhysioNet/CinC Challenge dataset (PCCD) and our collected publicly available pediatric heart sound dataset (PHSD) are used to evaluate the performance of the model by 10-fold cross-validation. The model in this paper achieves accuracies of 96.3 % and 99.1 % on these two datasets respectively, demonstrating its capability to reach the state-of-the-art level. We also open the code of our algorithms.

Novel K-Nearest Neighbor With Convolutional Neural Networks (KNN-CNN) For Accurate Brain Tumor Detection In Image Mining

Brain tumor classification plays a crucial role in early diagnosis and effective treatment planning. In this paper, we propose a novel approach, K-Nearest Neighbor with Convolutional Neural Networks (KNN-CNN), for accurate brain tumor classification. The proposed method combines the strengths of K-Nearest Neighbor (KNN) and Convolutional Neural Networks (CNNs) to leverage both traditional feature-based classification and deep learning-based feature extraction. We use CNNs to learn high-level features from brain tumor images, and KNN is employed to classify tumors based on the extracted features. The experimental results on a brain tumor dataset demonstrate the effectiveness and efficiency of the KNN-CNN approach, achieving high classification accuracy and outperforming traditional methods.

Classification of Fracture Risk in Fallers Using Dual-Energy X-Ray Absorptiometry (DXA) Images and Deep Learning-Based Feature Extraction.

Dual-energy X-ray absorptiometry (DXA) scans are one of the most frequently used imaging techniques for calculating bone mineral density, yet calculating fracture risk using DXA image features is rarely performed. The objective of this study was to combine deep neural networks, together with DXA images and patient clinical information, to evaluate fracture risk in a cohort of adults with at least one known fall and age-matched healthy controls. DXA images of the entire body as, well as isolated images of the hip, forearm, and spine (1488 total), were obtained from 478 fallers and 48 non-faller controls. A modeling pipeline was developed for fracture risk prediction using the DXA images and clinical data. First, self-supervised pretraining of feature extractors was performed using a small vision transformer (ViT-S) and a convolutional neural network model (VGG-16 and Resnet-50). After pretraining, the feature extractors were then paired with a multilayer perceptron model, which was used for fracture risk classification. Classification was achieved with an average area under the receiver-operating characteristic curve (AUROC) score of 74.3%. This study demonstrates ViT-S as a promising neural network technique for fracture risk classification using DXA scans. The findings have future application as a fracture risk screening tool for older adults at risk of falls. © 2023 The Authors. JBMR Plus published by Wiley Periodicals LLC on behalf of American Society for Bone and Mineral Research.

Teeth Lesion Detection Using Deep Learning and the Internet of Things Post-COVID-19.

With a view of the post-COVID-19 world and probable future pandemics, this paper presents an Internet of Things (IoT)-based automated healthcare diagnosis model that employs a mixed approach using data augmentation, transfer learning, and deep learning techniques and does not require physical interaction between the patient and physician. Through a user-friendly graphic user interface and availability of suitable computing power on smart devices, the embedded artificial intelligence allows the proposed model to be effectively used by a layperson without the need for a dental expert by indicating any issues with the teeth and subsequent treatment options. The proposed method involves multiple processes, including data acquisition using IoT devices, data preprocessing, deep learning-based feature extraction, and classification through an unsupervised neural network. The dataset contains multiple periapical X-rays of five different types of lesions obtained through an IoT device mounted within the mouth guard. A pretrained AlexNet, a fast GPU implementation of a convolutional neural network (CNN), is fine-tuned using data augmentation and transfer learning and employed to extract the suitable feature set. The data augmentation avoids overtraining, whereas accuracy is improved by transfer learning. Later, support vector machine (SVM) and the K-nearest neighbors (KNN) classifiers are trained for lesion classification. It was found that the proposed automated model based on the AlexNet extraction mechanism followed by the SVM classifier achieved an accuracy of 98%, showing the effectiveness of the presented approach.

Visual measurement of grinding surface roughness based on feature fusion

Due to the random surface texture and weak features of the grinding processing, most current visual measurement of grinding surface roughness is predicted by designing feature indicators, but its imaging environment is more demanding and the indicator design is more artificial. While deep learning can achieve feature self-extraction, the existing deep learning-based feature extraction uses a single convolutional model to extract features, which tends to make the extracted features noise, with low resolution and poor perception of details. To address the above problems, this paper proposes a feature fusion-based method for measuring grinding surface roughness. The method adopts different feature extraction ways and fuses the extracted features to obtain more features while also improving the generalization ability of the model, and verifies the adaptability of the model in different lighting environments. The experimental results show that deep learning self-extracted features based on feature fusion can effectively solve the problem of weak feature information on grinding surface roughness that is difficult to identify, and the model has high detection accuracy across different lighting environments, thus laying the foundation for the automated visual online measurement of grinding surface roughness.

Deep learning-based classification model for GPR151 activator activity prediction

BackgroundGPR151 is a kind of protein belonging to G protein-coupled receptor family that is closely associated with a variety of physiological and pathological processes.The potential use of GPR151 as a therapeutic target for the management of metabolic disorders has been demonstrated in several studies, highlighting the demand to explore its activators further. Activity prediction serves as a vital preliminary step in drug discovery, which is both costly and time-consuming. Thus, the development of reliable activity classification model has become an essential way in the process of drug discovery, aiming to enhance the efficiency of virtual screening.ResultsWe propose a learning-based method based on feature extractor and deep neural network to predict the activity of GPR151 activators. We first introduce a new molecular feature extraction algorithm which utilizes the idea of bag-of-words model in natural language to densify the sparse fingerprint vector. Mol2vec method is also used to extract diverse features. Then, we construct three classical feature selection algorithms and three types of deep learning model to enhance the representational capacity of molecules and predict activity label by five different classifiers. We conduct experiments using our own dataset of GPR151 activators. The results demonstrate high classification accuracy and stability, with the optimal model Mol2vec-CNN significantly improving performance across multiple classifiers. The svm classifier achieves the best accuracy of 0.92 and F1 score of 0.76 which indicates promising applications for our method in the field of activity prediction.ConclusionThe results suggest that the experimental design of this study is appropriate and well-conceived. The deep learning-based feature extraction algorithm established in this study outperforms traditional feature selection algorithm for activity prediction. The model developed can be effectively utilized in the pre-screening stage of drug virtual screening.

Hybrid Deep Learning and Discrete Wavelet Transform-Based ECG Biometric Recognition for Arrhythmic Patients and Healthy Controls.

The intrinsic and liveness detection behavior of electrocardiogram (ECG) signals has made it an emerging biometric modality for the researcher with several applications including forensic, surveillance and security. The main challenge is the low recognition performance with datasets of large populations, including healthy and heart-disease patients, with a short interval of an ECG signal. This research proposes a novel method with the feature-level fusion of the discrete wavelet transform and a one-dimensional convolutional recurrent neural network (1D-CRNN). ECG signals were preprocessed by removing high-frequency powerline interference, followed by a low-pass filter with a cutoff frequency of 1.5 Hz for physiological noises and by baseline drift removal. The preprocessed signal is segmented with PQRST peaks, while the segmented signals are passed through Coiflets' 5 Discrete Wavelet Transform for conventional feature extraction. The 1D-CRNN with two long short-term memory (LSTM) layers followed by three 1D convolutional layers was applied for deep learning-based feature extraction. These combinations of features result in biometric recognition accuracies of 80.64%, 98.81% and 99.62% for the ECG-ID, MIT-BIH and NSR-DB datasets, respectively. At the same time, 98.24% is achieved when combining all of these datasets. This research also compares conventional feature extraction, deep learning-based feature extraction and a combination of these for performance enhancement, compared to transfer learning approaches such as VGG-19, ResNet-152 and Inception-v3 with a small segment of ECG data.

SIGN LANGUAGE DETECTION USING ACTION RECOGNITION WITH PYTHON

Sign language detection plays a crucial role in bridging the communication gap between individuals with hearing impairments and the rest of the population. This study presents a novel approach for sign language detection using action recognition and LSTM deep learning models implemented in Python. The proposed system aims to accurately recognize and interpret sign language gestures in real-time. The research focuses on leveraging the power of deep learning techniques, specifically the Long Short-Term Memory (LSTM) architecture, to effectively capture temporal dependencies and classify sign language gestures. A comprehensive dataset comprising a wide range of sign language gestures is collected and preprocessed to train and evaluate the LSTM model. The pipeline for sign language detection consists of several stages, including video acquisition, preprocessing, feature extraction, and model training. In the pre-processing stage, the acquired video data is segmented into individual frames, and various image processing techniques are applied to enhance the quality and remove noise. Next, robust features are extracted from the preprocessed frames using techniques like optical flow or deep learning-based feature extraction methods. The LSTM model is then trained on the extracted features to learn the temporal dynamics of sign language gestures. Transfer learning is also explored to leverage pre-trained models on large-scale action recognition datasets. The trained model is evaluated using a comprehensive set of metrics, including accuracy, precision, recall, and F1-score, to assess its performance. Experimental results demonstrate the effectiveness of the proposed approach in accurately recognizing sign language gestures. The LSTM-based deep learning model achieves a high accuracy rate, showcasing its capability to handle the temporal nature of sign language. The system exhibits real-time performance, enabling it to be deployed in various applications, such as real-time interpretation tools or assistive devices for individuals with hearing impairments

Deep Learning-Based Feature Extraction from Whole-Body PET/CT Employing Maximum Intensity Projection Images: Preliminary Results of Lung Cancer Data.

Deep learning (DL) has been widely used in various medical imaging analyses. Because of the difficulty in processing volume data, it is difficult to train a DL model as an end-to-end approach using PET volume as an input for various purposes including diagnostic classification. We suggest an approach employing two maximum intensity projection (MIP) images generated by whole-body FDG PET volume to employ pre-trained models based on 2-D images. As a retrospective, proof-of-concept study, 562 [18F]FDG PET/CT images and clinicopathological factors of lung cancer patients were collected. MIP images of anterior and lateral views were used as inputs, and image features were extracted by a pre-trained convolutional neural network (CNN) model, ResNet-50. The relationship between the images was depicted on a parametric 2-D axes map using t-distributed stochastic neighborhood embedding (t-SNE), with clinicopathological factors. A DL-based feature map extracted by two MIP images was embedded by t-SNE. According to the visualization of the t-SNE map, PET images were clustered by clinicopathological features. The representative difference between the clusters of PET patterns according to the posture of a patient was visually identified. This map showed a pattern of clustering according to various clinicopathological factors including sex as well as tumor staging. A 2-D image-based pre-trained model could extract image patterns of whole-body FDG PET volume by using anterior and lateral views of MIP images bypassing the direct use of 3-D PET volume that requires large datasets and resources. We suggest that this approach could be implemented as a backbone model for various applications for whole-body PET image analyses.

Deep Learning-Based Feature Extraction with MRI Data in Neuroimaging Genetics for Alzheimer's Disease.

The prognosis and treatment of patients suffering from Alzheimer's disease (AD) have been among the most important and challenging problems over the last few decades. To better understand the mechanism of AD, it is of great interest to identify genetic variants associated with brain atrophy. Commonly, in these analyses, neuroimaging features are extracted based on one of many possible brain atlases with FreeSurf and other popular software; this, however, may cause the loss of important information due to our incomplete knowledge about brain function embedded in these suboptimal atlases. To address the issue, we propose convolutional neural network (CNN) models applied to three-dimensional MRI data for the whole brain or multiple, divided brain regions to perform completely data-driven and automatic feature extraction. These image-derived features are then used as endophenotypes in genome-wide association studies (GWASs) to identify associated genetic variants. When we applied this method to ADNI data, we identified several associated SNPs that have been previously shown to be related to several neurodegenerative/mental disorders, such as AD, depression, and schizophrenia.

Deep learning-based feature extraction coupled with multi class SVM for COVID-19 detection in the IoT era

Deep learning-based feature extraction coupled with multi class SVM for COVID-19 detection in the IoT era

Estimating Human Weight From a Single Image

Body weight, as one of the biometric traits, has been studied in both the forensic and medical domains. However, estimating weight directly from 2D images is particularly challenging since visual inspection is rather sensitive to the distance between the subject and camera, even for frontal view images. In this case, the widely used body mass index (BMI), which is associated with body height and weight, can be employed as a measure of weight to indicate health conditions. Previous works on the estimation of BMI have predominantly focused on using multiple 2D images, 3D images, or facial images; however, these cues are not always available. To address this issue, we explore the feasibility of obtaining BMI from a single 2D body image with the dual-branch regression framework proposed in this work. More specifically, the framework comprises an anthropometric feature computation branch and a deep learning-based feature extraction branch. One aggregation layer maps all the features to an estimated BMI value. In addition, a new public 2D image-to-BMI dataset, which contains 4189 images (1477 males and 2712 females) from approximately 3000 subjects with attributes including gender, age, height, and weight, was collected and released to facilitate the study. Extensive experiments confirm that the proposed framework combining anthropometric features and deep features outperforms the single-type feature approaches to BMI estimation in most cases.

Adaptive Weighted Flow Net Algorithm for Human Activity Recognition Using Depth Learned Features

Human Activity Recognition (HAR) from video data collections is the core application in vision tasks and has a variety of utilizations including object detection applications, video-based behavior monitoring, video classification, and indexing, patient monitoring, robotics, and behavior analysis. Although many techniques are available for HAR in video analysis tasks, most of them are not focusing on behavioral analysis. Hence, a new HAR system analysis the behavioral activity of a person based on the deep learning approach proposed in this work. The most essential aim of this work is to recognize the complex activities that are useful in many tasks that are based on object detection, modelling of individual frame characteristics, and communication among them. Moreover, this work focuses on finding out the human actions from various video resolutions, invariant human poses, and nearness of multi objects. First, we identify the key and essential frames of each activity using histogram differences. Secondly, Discrete Wavelet Transform (DWT) is used in this system to extract coefficients from the sequence of key-frames where the activity is localized in space. Finally, an Adaptive Weighted Flow Net (AWFN) algorithm is proposed in this work for effective video activity recognition. Moreover, the proposed algorithm has been evaluated by comparing it with the existing Visual Geometry Group (VGG-16) convolution neural networks for making performance comparisons. This work focuses on competent deep learning-based feature extraction to discriminate the activities for performing the classification accuracy. The proposed model has been evaluated with VGG-16 using a combination of regular UCF-101 activity datasets and also in very challenging Low-quality videos such as HMDB51. From these investigations, it is proved that the proposed AWFN approach gives higher detection accuracy of 96%. It is approximately 0.3% to 7.88% of higher accuracy than state-of-art methods.

MRNA Biomarkers for Invasive Breast Cancer based on a Deep Feature Selection Approach

Early detection of breast cancer and effective identification of its correct stage remain major challenges for healthcare professionals. Testing the tumour for Oestrogen Receptor and Progesterone Receptor is a standard part of the initial evaluation of breast cancer diagnosis and treatment planning. Several expression profiling studies have illustrated that the expression of these hormone receptors is linked with diverse genetic variations, which means that several mutated genes can a affect the development and progression of breast cancer and contribute to its heterogeneity. Unfortunately, due to the high dimensionality and low sample size nature of microarray data, traditional statistical feature selection techniques fail to identify genes that could act as risk factors for breast cancer. Inspired by this, we developed a deep learning-based feature extraction module with a weight interpretation method to select a subset of robust biomarkers across three different mRNA expression data sets from The Cancer Genome Atlas program (TCGA). For a discovered feature (a gene) to be accepted for further investigation, it must have been independently selected by the weight interpretation method from each of the deep feature extraction modules (each having been trained on a different data set). The small panel of discovered biomarkers was then subsequently evaluated using a range of classifiers to ascertain their predictive ability with respect to the above hormone receptor status. We observed strong evidence that the upregulation in the expression levels of highly positively weighted genes within the deep feature selection modules and the down regulation in the expression levels of the highly negatively weighted genes both indicated the strong likelihood of a patient experiencing ER+/PR+ invasive breast cancer. In addition, we discovered a number of potentially novel biomarkers worthy of further consideration.