scholarly journals Breast Segmentation and Probable Region Identification for Breast Cancer using DL-CNN

2019 ◽  
Vol 9 (2) ◽  
pp. 16-22
Keyword(s):  
Breast Cancer ◽  
False Positive ◽  
Research Work ◽  
Region Of Interest ◽  
Layered Architecture ◽  
Proposed Model ◽  
False Positive Reduction ◽  
Breast Segmentation ◽  
Novel Method ◽  
Result Analysis

Mammography is one of the key method used for detecting the breast cancer, several researcher has proposed the detection and segmentation method, however absolute solution has not developed till now and they have certain limitation and still it is one of the major challenge for finding the region in masses. Hence in this research work we have developed and design a novel method named as DL-CNN (Dual Layered) architecture CNN. The main intention of the model is segmentation and probable region identification. DL-CNN is based on the Convolution Neural Network. It has two layer first layer is applied for identifying the probable region whereas the second layer is used for segmentation and minimizing the false positive Reduction. In order to evaluate the DL-CNN algorithm by using the In Breast Dataset. Moreover the proposed model is compared against the various model in terms of ROI(Region of Interest), Dice_Index and False positive per Image. Result analysis shows that our model outperforms the existing

scholarly journals Computer aided diagnosis for breast cancer based on the gabor filter technique

2020 ◽  
Vol 10 (5) ◽  
pp. 5235 ◽  
Author(s):  
Mohammed Y. Kamil
Keyword(s):  
Breast Cancer ◽  
Gabor Filter ◽  
Diagnostic System ◽  
Region Of Interest ◽  
Histogram Equalization ◽  
Support Vector ◽  
Cancer Early Detection ◽  
False Positive Reduction ◽  
Computer Aided ◽  
Aided Diagnosis

The most prominent reason for the death of women all over the world is breast cancer. Early detection of cancer helps to lower the death rate. Mammography scans determine breast tumors in the first stage. As the mammograms have slight contrast, thus, it is a blur to the radiologist to recognize micro growths. A computer-aided diagnostic system is a powerful tool for understanding mammograms. Also, the specialist helps determine the presence of the breast lesion and distinguish between the normal area and the mass. In this paper, the Gabor filter is presented as a key step in building a diagnostic system. It is considered a sufficient method to extract the features. That helps us to avoid tumor classification difficulties and false-positive reduction. The linear support vector machine technique is used in this system for results classification. To improve the results, adaptive histogram equalization pre-processing procedure is employed. Mini-MIAS database utilized to evaluate this method. The highest accuracy, sensitivity, and specificity achieved are 98.7%, 98%, 99%, respectively, at the region of interest (30×30). The results have demonstrated the efficacy and accuracy of the proposed method of helping the radiologist on diagnosing breast cancer.

scholarly journals Ultrasound image segmentation through deep learning based improvised U-Net

2021 ◽  
Vol 21 (3) ◽  
pp. 1424
Author(s):  
Nayana R. Shenoy ◽  
Anand Jatti
Keyword(s):  
Image Segmentation ◽  
Thyroid Cancer ◽  
Thyroid Gland ◽  
Thyroid Nodule ◽  
Research Work ◽  
Ultrasound Image ◽  
Region Of Interest ◽  
Great Accuracy ◽  
Proposed Model ◽  
High Level

<p><span id="docs-internal-guid-cea63826-7fff-8080-83de-ad2ba4604953"><span>Thyroid nodule are fluid or solid lump that are formed within human’s gland and most thyroid nodule doesn’t show any symptom or any sign; moreover there are certain percentage of thyroid gland are cancerous and which could lead human into critical situation up to death. Hence, it is one of the important type of cancer and also it is important for detection of cancer. Ultrasound imaging is widely popular and frequently used tool for diagnosing thyroid cancer, however considering the wide application in clinical area such estimating size, shape and position of thyroid cancer. Further, it is important to design automatic and absolute segmentation for better detection and efficient diagnosis based on US-image. Segmentation of thyroid gland from the ultrasound image is quiet challenging task due to inhomogeneous structure and similar existence of intestine. Thyroid nodule can appear anywhere and have any kind of contrast, shape and size, hence segmentation process needs to designed carefully; several researcher have worked in designing the segmentation mechanism, however most of them were either semi-automatic or lack with performance metric, however it was suggested that U-Net possesses great accuracy. Hence, in this paper, we proposed improvised U-Net which focuses on shortcoming of U-Net, the main aim of this research work is to find the probable Region of interest and segment further. Furthermore, we develop High level and low-level feature map to avoid the low-resolution problem and information; later we develop dropout layer for further optimization. Moreover proposed model is evaluated considering the important metrics such as accuracy, Dice Coefficient, AUC, F1-measure and true positive; our proposed model performs better than the existing model. </span></span></p>

scholarly journals Computerized detection of lung nodules with an enhanced false positive reduction scheme

2021 ◽  
Author(s):  
Negar Memarian
Keyword(s):  
False Positive ◽  
State Of The Art ◽  
Lung Nodule ◽  
Region Of Interest ◽  
Hybrid Scheme ◽  
Rule Based ◽  
Lung Nodule Detection ◽  
False Positive Reduction ◽  
Nodule Detection ◽  
Interest Detection

This thesis is based on the original investigations of the author in the field of computerized lung nodule detection in computed tomography (CT) images. The methodologies discussed in this thesis include two main topics: region of interest detection and enhanced false positive (FP) reduction. The system, which is developed to be a supplementary diagnostic tool for radiologists, first spots all the regions suspected to be nodules in the lung. Then it pins down the candidates with the highest possibility of being nodules through a series of rule based filtering stages. Finally, an enhanced false positive reduction system, which is in fact designed as a hybrid scheme based on learning algorithms, reduces the false positive detections further. The overall system performs with 72% sensitivity and 2.42 FP/slice, which competes with state-of-the-art methods. The system was tested on a database consisting of 24 pediatric clinical subjects with 1190 images and 154 metastatic nodules.

scholarly journals Fusion of Random Projection, Multi-Resolution Features and Distance Weighted K Nearest Neighbor for Masses Detection in Mammographic Images

2018 ◽  
Vol 10 (3) ◽  
pp. 1030 ◽  
Author(s):  
Viet Dung Nguyen ◽  
Minh Dong Le
Keyword(s):  
Breast Cancer ◽  
Nearest Neighbor ◽  
Random Projection ◽  
K Nearest Neighbor ◽  
Detection Systems ◽  
Mammographic Images ◽  
Extraction Step ◽  
False Positive Reduction ◽  
Distance Weighted ◽  
Novel Method

<p>Breast cancer is the top cancer in women both in the developed and the developing world. For early detection of the disease, mammography is still the most effective method beside ultrasound and magnetic resonance imaging. Computer Aided Detection systems have been developed to aid radiologists in diagnosing breast cancer. Different methods were proposed to overcome the main drawback of producing large number of False Positives.  In this paper, we presented a novel method for masses detection in mammograms. To describe masses, multi-resolution features were utilized. In feature extraction step, we calculated multi-resolution Block Difference Inverse Probability features and multi-resolution statistical features. Once the descriptors were extracted, we deployed random projection and distance weighted K Nearest Neighbor to classify the detected masses. The result is quite sanguine with sensitivity, false positive reduction and time for carrying out the algorithm</p>

scholarly journals RESCNN: A Deep Learning Approach for Unmasking Face Mask

2021 ◽  
Vol 7 (1) ◽  
pp. 49-55
Author(s):  
Bhusra Fatima ◽  
Arun Kumar Jhapate
Keyword(s):  
Research Work ◽  
Face Mask ◽  
Surveillance Systems ◽  
The Public ◽  
Facial Region ◽  
Physical Conditions ◽  
Proposed Model ◽  
The Face ◽  
Current Scenario ◽  
Result Analysis

Due to outbreak of COVID-19 pandemic, the trend of wearing mask is rising all over the world. Before such pandemic people wear mask only to protect themselves from pollution. While other people are self-conscious about their looks, they hide their emotions from the public by hiding their faces. But in current scenario, after pandemic, it is compulsory to wear mask everywhere as researchers and doctors have proved that wearing face masks works on impeding COVID-19 transmission. Nowadays, all attendance system or surveillance systems, etc. are integrated with AI technology in which face recognition is considered as input variable. So, there is need to determine all facial landmarks to recognize an individual. In this research work, Residual Convolution Neural Network (ResCNN), network is designed and simulated which unmasks the face mask present on face and restore mask area and recognize an individual. The result analysis is performed in three different cases or scenario, one normal frontal facial region with mask, in another case the masked face is tilted and in third case the noisy masked face is taken as input. The noise in image occurs due to many physical conditions. The dataset for training of ResCNN is prepared by masking facial images taken from CelebA dataset and MFR datasets to prove the efficiency of the proposed model.

scholarly journals Computerized detection of lung nodules with an enhanced false positive reduction scheme

2021 ◽  
Author(s):  
Negar Memarian
Keyword(s):  
False Positive ◽  
State Of The Art ◽  
Lung Nodule ◽  
Region Of Interest ◽  
Hybrid Scheme ◽  
Rule Based ◽  
Lung Nodule Detection ◽  
False Positive Reduction ◽  
Nodule Detection ◽  
Interest Detection

This thesis is based on the original investigations of the author in the field of computerized lung nodule detection in computed tomography (CT) images. The methodologies discussed in this thesis include two main topics: region of interest detection and enhanced false positive (FP) reduction. The system, which is developed to be a supplementary diagnostic tool for radiologists, first spots all the regions suspected to be nodules in the lung. Then it pins down the candidates with the highest possibility of being nodules through a series of rule based filtering stages. Finally, an enhanced false positive reduction system, which is in fact designed as a hybrid scheme based on learning algorithms, reduces the false positive detections further. The overall system performs with 72% sensitivity and 2.42 FP/slice, which competes with state-of-the-art methods. The system was tested on a database consisting of 24 pediatric clinical subjects with 1190 images and 154 metastatic nodules.

Sign in / Sign up

Export Citation Format

Share Document