Evija Cibulska
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Evija Cibulska has a total of 12 co-authors. Here are the top co-authors by number of shared publications:
- Alexander Derjabo
Riga Eastern Univ. Hospital (Latvia)
1 shared publication - Alexey Lihachev
Institute of Atomic Physics and Spectroscopy, Faculty of Science and Technology, University of Latvia, LV-1004 Riga, Latvia
1 shared publication Andrey Bondarenko
Riga Technical Univ. (Latvia)
1 shared publicationDmitrijs Bliznuks
Riga Technical Univ. (Latvia)
1 shared publication- Dmitrijs Bliznuks
Institute of Applied Computer Systems, Riga Technical University, LV-1048 Riga, Latvia
1 shared publication - Emilija Vija Plorina
Institute of Atomic Physics and Spectroscopy, Faculty of Science and Technology, University of Latvia, Riga, Latvia, LV-1004
1 shared publication
Latest Publications
- 27 February 2024
- Published by Springer Nature
- p. 61-68
Collaborative Virtual Reality Environment Structural Model Development for Higher Education Remote Learning
- 27 February 2024
- Published by Springer Nature
- p. 61-68
Collaborative Virtual Reality Environment Structural Model Development for Higher Education Remote Learning
- 28 October 2020
- Published by SPIE-Intl Soc Optical Eng
- Vol. 11585, 15
Challenges of automatic processing of large amount of skin lesion multispectral data
This work will describe the challenges involved in setting up automatic processing for a large differentiated data set. In this study, a multispectral (skin diffuse reflection images using 526nm (green), 663nm (red), and 964nm (infrared) illumination and autofluorescence (AF) image using 405 nm excitation) data set with 756 lesions (3024 images) was processed. Previously, using MATLAB software, finding markers, correctly segmenting images with dark edges and image alignment were the main causes of the problems in automatic data processing. To improve automatic processing and eliminate the use of licensed software, the latter was substituted with the open source Python environment. For more precise segmentation of skin markers and skin lesions, as well for image alignment, the processing of artificial neural networks was utilized. The resulting processing method solves most of the issues of the MATLAB script. However, for even more accurate results, it is necessary to provide more accurate ground-truth segmentation masks and generate more input data to increase the training image database by using data augmentation.
- 28 October 2020
- Published by SPIE-Intl Soc Optical Eng
- Vol. 11585, 15
Challenges of automatic processing of large amount of skin lesion multispectral data
This work will describe the challenges involved in setting up automatic processing for a large differentiated data set. In this study, a multispectral (skin diffuse reflection images using 526nm (green), 663nm (red), and 964nm (infrared) illumination and autofluorescence (AF) image using 405 nm excitation) data set with 756 lesions (3024 images) was processed. Previously, using MATLAB software, finding markers, correctly segmenting images with dark edges and image alignment were the main causes of the problems in automatic data processing. To improve automatic processing and eliminate the use of licensed software, the latter was substituted with the open source Python environment. For more precise segmentation of skin markers and skin lesions, as well for image alignment, the processing of artificial neural networks was utilized. The resulting processing method solves most of the issues of the MATLAB script. However, for even more accurate results, it is necessary to provide more accurate ground-truth segmentation masks and generate more input data to increase the training image database by using data augmentation.
- 28 October 2020
- Published by SPIE-Intl Soc Optical Eng
- Vol. 11585, 16
Deep learning model deploying on embedded skin cancer diagnostic device
The number of research papers, where neural networks are applied in medical image analysis is growing. There is a proof that Convolutional Neural Networks (CNN) are able to differentiate skin cancer from nevi with greater accuracy than experienced specialists on average (sensitivity 82% and 73% accordingly).1 Team's latest research2 allows achieving even greater accuracy, by using specific narrow-band illumination. Nevertheless, the overall probability of early skin cancer detection depends on the availability of diagnostic tools. If screening tools will be available to a high number of general practices, the chance of disease detection will increase. The previous research3 shows that scalable cloud service is able to process a high number of users. After a certain number of users, the overall cost of the system, including cloud processing expenses and cost of high computational power portable device, might be higher if compared to an on-premises solution, where each device is capable of diagnosing without Internet access. It might be cheaper to equip devices with additional neural processing unit (NPU) and exclude cloud processing. Another option is to make screening available by using the newest smartphones that are equipped with NPU.4 The problem of using the NPU is that they are limited in storage space, accuracy, and features. Therefore, a full-size CNN model should be adapted and minimized to fit in a limited NPU. Research reviews existing CNN optimization methods and proposes the most accurate for skin cancer diagnostics. The paper evaluates CNN prediction losses when the model's elements’ precision is reduced from 32 bits to 8 and rounded to integer values.
- 28 October 2020
- Published by SPIE-Intl Soc Optical Eng
- Vol. 11585, 16
Deep learning model deploying on embedded skin cancer diagnostic device
The number of research papers, where neural networks are applied in medical image analysis is growing. There is a proof that Convolutional Neural Networks (CNN) are able to differentiate skin cancer from nevi with greater accuracy than experienced specialists on average (sensitivity 82% and 73% accordingly).1 Team's latest research2 allows achieving even greater accuracy, by using specific narrow-band illumination. Nevertheless, the overall probability of early skin cancer detection depends on the availability of diagnostic tools. If screening tools will be available to a high number of general practices, the chance of disease detection will increase. The previous research3 shows that scalable cloud service is able to process a high number of users. After a certain number of users, the overall cost of the system, including cloud processing expenses and cost of high computational power portable device, might be higher if compared to an on-premises solution, where each device is capable of diagnosing without Internet access. It might be cheaper to equip devices with additional neural processing unit (NPU) and exclude cloud processing. Another option is to make screening available by using the newest smartphones that are equipped with NPU.4 The problem of using the NPU is that they are limited in storage space, accuracy, and features. Therefore, a full-size CNN model should be adapted and minimized to fit in a limited NPU. Research reviews existing CNN optimization methods and proposes the most accurate for skin cancer diagnostics. The paper evaluates CNN prediction losses when the model's elements’ precision is reduced from 32 bits to 8 and rounded to integer values.