Transfer learning with GoogLeNet for detection of lung cancer

Abstract

<p class="p1">The use of computer algorithms has gained momentum in filling/assisting roles of specialists especially in early diagnosis scenarios. This paper proposes the employment of deep neural networks (DNN) to detect images with malignant nodules of lung computed tomography (CT). The method includes subjecting input images to a simple and fast pre-processing which isolates regions of interest (ROI), that’s the lungs dominated area, ridding the images of other surrounding tissues and artefacts. Centered and size normalized images are then fed to a deep neural network for training and validation. In this work transfer learning is used to readjust GoogLeNet DNN to learn this medical data. This includes allowing final layers of the DNN to evolve while restricting deep layers. In this setting, a rough, unprocessed dataset, the IQ-OTH/NCCD lung cancer dataset was used to train/validate the proposed algorithm. Experimental results show that this algorithm scores 94.38% accuracy, which outperforms benchmark method previously used with this dataset.</p>

A convolutional neural network framework for classifying inappropriate online video contents

“…According to a literature survey [32] a transfer learning method can currently be adopted to improve the performance of traditional machine learning methods by transferring information from a related domain. There are many applications that successfully applied transfer learning to enhance the model performance, e.g., sports video classification using pre-trained neural network [33], classification of lung cancer using pre-trained convolutional neural networks [34], automated fruit recognition using pre-trained models [35], and plant leaf disease classification using a pre-trained model [36]. As suggested in [37], for improving the CNN model with the tranfer learning technique, a pre-trained model, namely EfficientNet-B0, which is suitable with the dataset used in this study (image size of 224x224 pixel), was used.…”

“…In transfer learning, the knowledge of existing trained models is transferred to build new classification models. There are several research studies that investigated on classification performance when using transfer learning, e.g., sports video classification using pre-trained neural network [33], classification of lung cancer based on CNN and transfer learning with GoogLeNet [34], fruit recognition using pre-trained models [35], and leaf disease classification based on a pre-trained model [36]. In research Ramesh and Mahesh [33], reported the comparison performance results between a CNN model and a pre-trained CNN model for classifying sports categories from video collected from YouTube.…”

“…The experimental results demonstrated that the pre-trained CNN model performed higher accuracy than the CNN model, with the accuracy of 92.68% and 88.75%, respectively. AL-Huseiny and Sajit [34] developed a CNN classification model using transfer learning with GoogLeNet for detecting lung cancer. The model was trained and validated by using computed tomography (CT) scan images.…”

A modified residual network for detection and classification of Alzheimer’s disease

“…Researchers have demonstrated a promising alternative method known as transfer learning to overcome these obstacles. Transfer learning means improving the learning of a new task through transferring knowledge from a previously learned task [25]. The fundamental goal of this research is to compare the performance of pre-trained ResNet50 against the modified residual network in detecting and automatically classifying Alzheimer's disease using MRI scans.…”

Breast Cancer Cad System by Using Transfer Learning and Enhanced Roi

“…This is an important area of research because during conventional examination by a radiologist, non-cancerous lesions can be misclassified as a cancer (false-positive), while malignancies may be missed (false-negative) resulting in radiologists failing to detect between 10% and 30% of breast cancers cases. Computer aided diagnosis (CAD) is aimed to speed up and enhance the diagnosis process via assisting specialists in the detection and classification of the breast cancer by deploying scalable computerized diagnostic tools, hence, restricting the occurrence of human related shortcomings (AL-Huseiny & Sajit, 2021;Batra et al, 2020;Jalalian et al, 2013;Tan, Sim & Ting, 2017). The introduction of computerized approaches into medical procedures is highly reliant on the sensing mechanisms.…”