Wyniki wyszukiwania - BazTech

1

Quantitative impact of label noise on the quality of segmentation of brain tumors on MRI scans

Marcinkiewicz Michał, Mrukwa Grzegorz

Annals of Computer Science and Information Systems

|

2019

|

Vol. 18

61--65

EN

Over the last few years, deep learning has proven to be a great solution to many problems, such as image or text classification. Recently, deep learning-based solutions have outperformed humans on selected benchmark datasets, yielding a promising future for scientific and real-world applications. Training of deep learning models requires vast amounts of high quality data to achieve such supreme performance. In real-world scenarios, obtaining a large, coherent, and properly labeled dataset is a challenging task. This is especially true in medical applications, where high-quality data and annotations are scarce and the number of expert annotators is limited. In this paper, we investigate the impact of corrupted ground-truth masks on the performance of a neural network for a brain tumor segmentation task. Our findings suggest that a) the performance degrades about 8% less than it could be expected from simulations, b) a neural network learns the simulated biases of annotators, c) biases can be partially mitigated by using an inversely-biased dice loss function.

2

Wpływ odkrycia promieniotwórczości na rozwój diagnostyki i leczenia nowotworów

Kułakowski A.

Analityka : nauka i praktyka

|

2011

|

nr 3

4-9

3

Zapach choroby. Czy pies może być detektorem?

Rudnicka J., Ligor T., Jezierski T., Buszewski B.

Analityka : nauka i praktyka

|

2010

|

nr 2

15-20

4

Zastosowanie termowizji w badaniach przesiewowych na przykładzie zmian nowotworowych piersi

Więcek M., Strąkowski R., Jakubowska T., Więcek B.

Acta Bio-Optica et Informatica Medica. Inżynieria Biomedyczna

|

2010

|

Vol. 16, nr 1

49-55

PL

W pracy przedstawiono procedurę termowizyjnych badań przesiewowych, w których dokonano selekcji oraz klasyfikacji obrazów za pomocą sieci neuronowych. Metodę zastosowano do badań zmian nowotworowych piersi. Sieć nauczono za pomocą 15 termogramów obrazujących zmiany patologiczne i 15 termogramów z grupy kontrolnej w projekcji przedniej i bocznej. Dodatkowo zastosowano analizę różnicy cech piersi zdrowej i chorej, co zapewniło największą skuteczność klasyfikacji na poziomie 70%. Lepsze wyniki klasyfikacji uzależnione są od wielkości zbioru uczącego.

EN

In this work, the procedure of screening breast pathology based on thermovision examinations was presented. The thermal image features were evaluated and classified by means of neural networks. 15 thermal images representing pathological lesions and 15 images from control group were examined, both from fivntal and side positions of a patient. In addition, the differences of features of neoplastic and healthy breasts tissue, were evaluated. The classification efficiency up to 70% was achieved. The better classification can be obtained in case of bigger number of thermal images from both groups: pathological and control ones.

5

Stary problem a nowe fakty - wykorzystanie monitoringu a diagnostyce chorób nowotworowych

Buszewski B., Ulanowska A., Amann A.

Analityka : nauka i praktyka

|

2008

|

nr 3

18-23

6

Numerical simulation of the freezing process in biological tissue with tumor

Dziewoński M., Paruch M.

Zeszyty Naukowe Katedry Mechaniki Stosowanej / Politechnika Śląska

|

2006

|

z. 26

87-92

EN

In the paper the numerical simulation of freezing process of tissue with tumor is presented. In particular the action of external cylindrical cryoprobe is analyzed. On the stage of numerical simulations the finite element method (FEM) is applied. The computations have been done using the MSC MARC/MENTAT code supplemented by additional procedures taking into account the cyclic boundary conditions and temperature-dependent thermophysical parameters of the tissue and tumor [2].

7

Piąta Multidisciplinarna Konferencja Nauki o Leku

Przemysł Chemiczny

|

2006

|

T. 85, nr 6

435-437

8

Cancer of the oral epithelium located in the upper gingiva in the light of research of energy dispersive X-ray analysis method : first report

Cieślik T., Cieślik A., Namysłowska A., Sabat D., Męcik T., Szczurek Z.

Acta of Bioengineering and Biomechanics

|

2001

|

Vol. 3, nr 2

81-86

EN

A phenomenon of cancerous changes is one of the most difficult diagnostic and therapeutic challenges of contemporary medicine. In this paper, we have described the method which can contribute to further development of cooperation between medical doctors and physicists. We have analyzed 6 cases of cancer located in the upper gingiva using EDX method. All patients (4 males and 2 females) were diagnosed in the 1st Department and Clinic of Maxillofacial Surgery, Silesian Medical Academy in Zabrze. The authors observed the changes in the quality and the quantity of mineral composition of healthy and cancerous oral epithelia. The size of oral epithelium carcinoma in the clinical diagnosis was essentially connected with the concentration of calcium which developed into visible forms. This research will be continued, firstly, because of a small group of the patients examined and secondly, because the observed concentrations of calcium, sulphur and phosphorus have been substantially higher in the group of patients with advanced clinical forms of the cancer. We think that this physical method could help in early diagnostics and monitoring of the cancer located in the upper gingiva.