Wyniki wyszukiwania - BazTech

1

Breast cancer diagnosis: A systematic review

Wen Xin, Guo Xing, Wang Shuihua, Lu Zhihai, Zhang Yudong

Biocybernetics and Biomedical Engineering

|

2024

|

Vol. 44, no. 1

119--148

EN

The second-leading cause of death for women is breast cancer. Consequently, a precise early diagnosis is essential. With the rapid development of artificial intelligence, computer-aided diagnosis can efficiently assist radiologists in diagnosing breast problems. Mammography images, breast thermal images, and breast ultrasound images are the three ways to diagnose breast cancer. The paper will discuss some recent developments in machine learning and deep learning in three different breast cancer diagnosis methods. The three components of conventional machine learning methods are image preprocessing, segmentation, feature extraction, and image classification. Deep learning includes convolutional neural networks, transfer learning, and other methods. Additionally, the benefits and drawbacks of different methods are thoroughly contrasted. Finally, we also provide a summary of the challenges and potential futures for breast cancer diagnosis.

2

Multi-stage fully convolutional network for precise prostate segmentation in ultrasound images

Feng Yujie, Atabansi Chukwuemeka Clinton, Nie Jing, Liu Haijun, Zhou Hang, Zhao Huai, Hong Ruixia, Li Fang, Zhou Xichuan

Biocybernetics and Biomedical Engineering

|

2023

|

Vol. 43, no. 3

586--602

EN

Prostate cancer is one of the most commonly diagnosed non-cutaneous malignant tumors and the sixth major cause of cancer-related death generally found in men globally. Automatic segmentation of prostate regions has a wide range of applications in prostate cancer diagnosis and treatment. It is challenging to extract powerful spatial features for precise prostate segmentation methods due to the wide variation in prostate size, shape, and histopathologic heterogeneity among patients. Most of the existing CNN-based architectures often produce unsatisfactory results and inaccurate boundaries in prostate segmentation, which are caused by inadequate discriminative feature maps and the limited amount of spatial information. To address these issues, we propose a novel deep learning technique called Multi-Stage FCN architecture for 2D prostate segmentation that captures more precise spatial information and accurate prostate boundaries. In addition, a new prostate ultrasound image dataset known as CCH-TRUSPS was collected from Chongqing University Cancer Hospital, including prostate ultrasound images of various prostate cancer architectures. We evaluate our method on the CCH-TRUSPS dataset and the publicly available Multi-site T2-weighted MRI dataset using five commonly used metrics for medical image analysis. When compared to other CNN-based methods on the CCH-TRUSPS test set, our Multi-Stage FCN achieves the highest and best binary accuracy of 99.15%, the DSC score of 94.90%, the IoU score of 89.80%, the precision of 94.67%, and the recall of 96.49%. The statistical and visual results demonstrate that our approach outperforms previous CNN-based techniques in all ramifications and can be used for the clinical diagnosis of prostate cancer.

3

Ultrasound spine image segmentation using multi-scale feature fusion Skip-Inception U-Net (SIU-Net)

Banerjee Sunetra, Lyu Juan, Huang Zixun, Leung Frank H.F, Lee Timothy, Yang De, Su Steven, Zheng Yongping, Ling Sai Ho

Biocybernetics and Biomedical Engineering

|

2022

|

Vol. 42, no. 1

341--361

EN

Scoliosis is a 3D spinal deformation where the spine takes a lateral curvature, forming an angle in the coronal plane. Diagnosis of scoliosis requires periodic detection, and frequent exposure to radiative imaging may cause cancer. A safer and more economical alternative imaging, i.e., 3D ultrasound imaging modality, is being explored. However, unlike other radiative modalities, an ultrasound image is noisy, which often suppresses the image’s useful information. Through this research, a novel hybridized CNN architecture, multi-scale feature fusion Skip-Inception U-Net (SIU-Net), is proposed for a fully automatic bony feature detection, which can be further used to assess the severity of scoliosis safely and automatically. The proposed architecture, SIU-Net, incorporates two novel features into the basic U-Net architecture: (a) an improvised Inception block and (b) newly designed decoder-side dense skip pathways. The proposed model is tested on 109 spine ultrasound image datasets. The architecture is evaluated using the popular (i) Jaccard Index (ii) Dice Coefficient and (iii) Euclidean distance, and compared with (a) the basic U-net segmentation model, (b) a more evolved UNet++ model, and (c) a newly developed MultiResUNet model. The results show that SIU-Net gives the clearest segmentation output, especially in the important regions of interest such as thoracic and lumbar bony features. The method also gives the highest average Jaccard score of 0.781 and Dice score of 0.883 and the lowest histogram Euclidean distance of 0.011 than the other three models. SIU-Net looks promising to meet the objectives of a fully automatic scoliosis detection system.

4

A Study on Thyroid Nodule Image Classification System Using Small Amount of Training Samples

Hoan Quoc Tuan, Pham Xuan Hien, Trinh Xuan Thang

Annals of Computer Science and Information Systems

|

2022

|

Vol. 33

129--133

EN

To reduce errors caused by traditional diagnostic methods that rely heavily on physician experience, the diagnostic systems based on computer-aided have been researched and developed to assist physicians in diagnosing thyroid disease. Therefore, performance of the computer systems plays an important role to improve the quality of diagnostic processes. Although there has been a number of publish related to this issue, those studies still have limitations in which needing large data sets for training classification models is considered the most concerning limitation of previous studies. To solve this limitations, in this work, a classification method using artificial intelligence with a small amount of data to analyze thyroid ultrasound images was proposed. Thus we can save time and effort for data collection and the classification model processing time. Through baseline tests with an open data set, especially thyroid digital image database (TDID), the proposed method has improved the limitations of previous methods.

5

A deep learning based approach for classification of abdominal organs using ultrasound images

Santhosh Reddy D., Rajalakshmi P., Mateen M. A.

Biocybernetics and Biomedical Engineering

|

2021

|

Vol. 41, no. 2

779--791

EN

Ultrasound imaging is one of the primary modalities used for diagnosing a multitude of medical conditions affecting organs and soft tissues the body. Unlike X-rays, which use ionizing radiation, ultrasound imaging utilizes non-hazardous acoustic waves and is widely preferred by doctors. However, ultrasound imaging sometimes requires substantial manual effort in the identification of organs during real-time scanning. Also, it is a challenging task if the scanning performed by an unskilled clinician does not comprise adequate information about the organ, leading to an incorrect diagnosis and thereby fatal consequences. Hence, the automated organ classification in such scenarios can offer potential benefits. In this paper, We propose a convolutional neural network-based architecture (CNNs), precisely, a transfer learning approach using ResNet, VGG, GoogleNet, and Inception models for accurate classification of abdominal organs namely kidney, liver, pancreas, spleen, and urinary bladder. The performance of the proposed framework is analyzed using in-house developed dataset comprising of 1906 ultrasound images. Performance analysis shows that the proposed framework achieves a classification accuracy and F1 score of 98.77% and 98.55%, respectively, on an average. Also, we provide the performance of the proposed architecture in comparison with the state-of-the-art studies.

6

Granular filter in medical image noise suppression and edge preservation

Wieclawek Wojciech, Pietka Ewa

Biocybernetics and Biomedical Engineering

|

2019

|

Vol. 39, no. 1

1--16

EN

An alternative non-linear filtering technique for medical image denoising while preserving edge is introduced. Two different variants of the approach i.e. crisp and fuzzy are developed. The solution is demonstrated based on US breast images as well as CT studies and gave promising results in comparison with commonly known and popular filtering techniques (i.e. spatial averaging and median, bilateral filter, anisotropic diffusion). Many different measures were used to evaluate the method. There are pixel-to-pixel error measures, structural information factors and edge preservation measures. The benefits are noticeable in all three categories.

7

Comparative assessment of texture features for the identification of cancer in ultrasound images: a review

Faust O., Acharya U. R., Meiburger K. M., Molinari F., Koh J. E. W., Yeong C. H., Kongmebhol P., Ng K. H.

Biocybernetics and Biomedical Engineering

|

2018

|

Vol. 38, no. 2

275--296

EN

In this paper, we review the use of texture features for cancer detection in Ultrasound (US) images of breast, prostate, thyroid, ovaries and liver for Computer Aided Diagnosis (CAD) systems. This paper shows that texture features are a valuable tool to extract diagnostically relevant information from US images. This information helps practitioners to discriminate normal from abnormal tissues. A drawback of some classes of texture features comes from their sensitivity to both changes in image resolution and grayscale levels. These limitations pose a considerable challenge to CAD systems, because the information content of a specific texture feature depends on the US imaging system and its setup. Our review shows that single classes of texture features are insufficient, if considered alone, to create robust CAD systems, which can help to solve practical problems, such as cancer screening. Therefore, we recommend that the CAD system design involves testing a wide range of texture features along with features obtained with other image processing methods. Having such a compet-itive testing phase helps the designer to select the best feature combination for a particular problem. This approach will lead to practical US based cancer detection systems which deliver real benefits to patients by improving the diagnosis accuracy while reducing health care cost.

8

Automated Detecting Symptoms Of Selected Gallbladder Illness Based On A Static Ultrasound Images Analysis

Bodzioch S.

Bio-Algorithms and Med-Systems

|

2006

|

Vol. 2, no. 3

33--42

EN

Automating a process of diagnosis always comes down to developing algorithms used to analyze the object of such diagnosis and verify the occurrence of symptoms related to a given affliction. Usually the final stage is to make a diagnosis based on the detected symptoms. This last stage can be carried out through either expert systems or a more classic approach, that is with the application of minimal distance methods. In the case of gallbladder diagnostics, the most important examination still remains image ultrasound diagnostics. This paper discusses the algorithms of a computer analysis of static ultrasound gallbladder images and detecting symptoms of the most frequent illness of this organ.

PL

Zautomatyzowanie procesu diagnostycznego zawsze sprowadza się do opracowania algorytmów analizujących przedmiot diagnozy oraz weryfikujących obecność symptomów związanych z danym schorzeniem. Ostatnim etapem jest postawienie diagnozy na podstawie wykrytych symptomów. Etap ten zwykle można zrealizować za pomocą systemów ekspertowych bądź podejściem bardziej klasycznym stosując metody minimalno-odległościowe. W przypadku diagnostyki pęcherzyka żółciowego, najważniejszym badaniem wciąż jest Obrazowa diagnostyka ultrasonograficzna. W niniejszym artykule zostaną omówione algorytmy komputerowej analizy statycznych obrazów USG pęcherzyka żółciowego oraz wykrywania symptomów dla najczęściej występujących schorzeń tego organu.