Jolanta Kawulok
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Jolanta Kawulok has a total of 14 co-authors. Here are the top co-authors by number of shared publications:
- Michal Kawulok
Department of Algorithmics and Software, Silesian University of Technology, 44-100 Gliwice, Poland
14 shared publications - Bogdan Smolka
Silesian University of Technology
5 shared publications - Jakub Nalepa
Silesian University of Technology
5 shared publications - Sebastian Deorowicz
Department of Algorithmics and Software, Silesian University of Technology, Gliwice, Poland
3 shared publications - Anna Stachowicz
Department of Algorithmics and Software, Silesian University of Technology, 44-100 Gliwice, Poland
1 shared publication Jakub Adler
Faculty of Automatic Control, Electronics and Computer Science, Silesian University of Technology
1 shared publication
Latest Publications
- Open Access
- 12 August 2026
- Published by MDPI AG in Applied Sciences
- Vol. 16 (16) , 8034
Uncertainty-Aware Stratified Scribble Propagation for Weakly Supervised Human Skin Segmentation
Segmenting human skin regions in digital images is a well-studied computer vision problem with a wide range of applications. State-of-the-art deep learning models achieve high accuracy, but their training requires large sets of precisely annotated pixel-level masks, whose preparation remains a time-consuming and tedious task. In this paper, we propose a weakly supervised framework that trains competitive skin segmentation networks from sparse, single-class scribble annotations. We exploit the concept of a well-established skin-aware shortest-path propagation scheme, adapting it to generate tri-level pseudo-labels consisting of high-confidence skin and background regions, and an uncertain region that is excluded from the training loss. We evaluate the framework on the ECU benchmark using two architectures: the lightweight Skinny convolutional neural network and the transformer-based SegFormer-B2. The proposed method achieves an F-score of 0.85 on Skinny using the reproducible simulated-scribble protocol and 0.87 using manually drawn scribbles. When the model is additionally fine-tuned on just 80 precisely annotated images, F-score improves to 0.89, approaching fully supervised accuracy while requiring substantially fewer dense annotations. The proposed pseudo-labeling framework also outperforms a content-agnostic ScribbleSup baseline, and the ECU-trained models transfer to the HGR hand-gesture dataset without any adaptation. While we evaluate the framework on skin segmentation, its underlying principles may be applicable beyond this specific task and point to broader potential in weakly supervised segmentation.
- 12 August 2026
- Published by MDPI AG in Applied Sciences
- Vol. 16 (16) , 8034
Open AccessUncertainty-Aware Stratified Scribble Propagation for Weakly Supervised Human Skin Segmentation
Segmenting human skin regions in digital images is a well-studied computer vision problem with a wide range of applications. State-of-the-art deep learning models achieve high accuracy, but their training requires large sets of precisely annotated pixel-level masks, whose preparation remains a time-consuming and tedious task. In this paper, we propose a weakly supervised framework that trains competitive skin segmentation networks from sparse, single-class scribble annotations. We exploit the concept of a well-established skin-aware shortest-path propagation scheme, adapting it to generate tri-level pseudo-labels consisting of high-confidence skin and background regions, and an uncertain region that is excluded from the training loss. We evaluate the framework on the ECU benchmark using two architectures: the lightweight Skinny convolutional neural network and the transformer-based SegFormer-B2. The proposed method achieves an F-score of 0.85 on Skinny using the reproducible simulated-scribble protocol and 0.87 using manually drawn scribbles. When the model is additionally fine-tuned on just 80 precisely annotated images, F-score improves to 0.89, approaching fully supervised accuracy while requiring substantially fewer dense annotations. The proposed pseudo-labeling framework also outperforms a content-agnostic ScribbleSup baseline, and the ECU-trained models transfer to the HGR hand-gesture dataset without any adaptation. While we evaluate the framework on skin segmentation, its underlying principles may be applicable beyond this specific task and point to broader potential in weakly supervised segmentation.
- 9 December 2022
- Published by Springer Nature
- p. 329-342
Toward Understanding the Impact of Input Data for Multi-Image Super-Resolution
- 9 December 2022
- Published by Springer Nature
- p. 329-342
Toward Understanding the Impact of Input Data for Multi-Image Super-Resolution
- 1 January 2020
- Published by INSTICC
- p. 218-225
Metagenomic Clustering in Search of Common Origin
- Open Access
- 13 November 2019
- Published by Springer Nature in Biology Direct
- Vol. 14 (1) , 1-23
Environmental metagenome classification for constructing a microbiome fingerprint
Background: Nowadays, not only are single genomes commonly analyzed, but also metagenomes, which are sets of, DNA fragments (reads) derived from microbes living in a given environment. Metagenome analysis is aimed at extracting crucial information on the organisms that have left their traces in an investigated environmental sample.In this study we focus on the MetaSUB Forensics Challenge (organized within the CAMDA 2018 conference) which consists in predicting the geographical origin of metagenomic samples. Contrary to the existing methods for environmental classification that are based on taxonomic or functional classification, we rely on the similarity between a sample and the reference database computed at a reads level.Results: We report the results of our extensive experimental study to investigate the behavior of our method and its sensitivity to different parameters. In our tests, we have followed the protocol of the MetaSUB Challenge, which allowed us to compare the obtained results with the solutions based on taxonomic and functional classification.Conclusions: The results reported in the paper indicate that our method is competitive with those based on taxonomic classification. Importantly, by measuring the similarity at the reads level, we avoid the necessity of using large databases with annotated gene sequences. Hence our main finding is that environmental classification of metagenomic data can be proceeded without using large databases required for taxonomic or functional classification.Reviewers: This article was reviewed by Eran Elhaik, Alexandra Bettina Graf, Chengsheng Zhu, and Andre Kahles.
- 13 November 2019
- Published by Springer Nature in Biology Direct
- Vol. 14 (1) , 1-23
Open AccessEnvironmental metagenome classification for constructing a microbiome fingerprint
Background: Nowadays, not only are single genomes commonly analyzed, but also metagenomes, which are sets of, DNA fragments (reads) derived from microbes living in a given environment. Metagenome analysis is aimed at extracting crucial information on the organisms that have left their traces in an investigated environmental sample.In this study we focus on the MetaSUB Forensics Challenge (organized within the CAMDA 2018 conference) which consists in predicting the geographical origin of metagenomic samples. Contrary to the existing methods for environmental classification that are based on taxonomic or functional classification, we rely on the similarity between a sample and the reference database computed at a reads level.Results: We report the results of our extensive experimental study to investigate the behavior of our method and its sensitivity to different parameters. In our tests, we have followed the protocol of the MetaSUB Challenge, which allowed us to compare the obtained results with the solutions based on taxonomic and functional classification.Conclusions: The results reported in the paper indicate that our method is competitive with those based on taxonomic classification. Importantly, by measuring the similarity at the reads level, we avoid the necessity of using large databases with annotated gene sequences. Hence our main finding is that environmental classification of metagenomic data can be proceeded without using large databases required for taxonomic or functional classification.Reviewers: This article was reviewed by Eran Elhaik, Alexandra Bettina Graf, Chengsheng Zhu, and Andre Kahles.
- 1 January 2018
- Published by INSTICC
- p. 182-187
Environmental Metagenome Classification for Soil-based Forensic Analysis
Metagenome analysis makes it possible to extract essential information on the organisms that have left their traces in a given environmental sample. In some cases, it is sufficient to determine the origin of an environmental sample, rather than being able to accurately identify the organisms living there (which may be a challenging task). For example, in forensic soil analysis, it could be possible to confirm or exclude that a defendant was present in a certain place by comparing a soil sample acquired from his belongings against the samples derived from a variety of places (including the suspected place). In this paper, we present a method to identify the environmental origins of metagenomic reads by comparing them with entire metagenomic collections derived from reference samples. For this purpose, we exploit our CoMeta program, which allows for fast classification of metagenome samples, and we apply it to classify the extracted soil metagenomes to various collections of so
- 1 January 2018
- Published by INSTICC
- p. 182-187
Environmental Metagenome Classification for Soil-based Forensic Analysis
Metagenome analysis makes it possible to extract essential information on the organisms that have left their traces in a given environmental sample. In some cases, it is sufficient to determine the origin of an environmental sample, rather than being able to accurately identify the organisms living there (which may be a challenging task). For example, in forensic soil analysis, it could be possible to confirm or exclude that a defendant was present in a certain place by comparing a soil sample acquired from his belongings against the samples derived from a variety of places (including the suspected place). In this paper, we present a method to identify the environmental origins of metagenomic reads by comparing them with entire metagenomic collections derived from reference samples. For this purpose, we exploit our CoMeta program, which allows for fast classification of metagenome samples, and we apply it to classify the extracted soil metagenomes to various collections of so
- 13 July 2016
- Published by SAGE Publications in Intelligent Data Analysis: An International Journal
- Vol. 20 (s1) , S121-S139
Hybrid adaptation for detecting skin in color images
It has been reported in many works on skin detection and segmentation from color images that skin color models suffer from low specificity and high variance of the skin color, and this problem can be addressed by conforming the skin model to a presen
- 13 July 2016
- Published by SAGE Publications in Intelligent Data Analysis: An International Journal
- Vol. 20 (s1) , S121-S139
Hybrid adaptation for detecting skin in color images
It has been reported in many works on skin detection and segmentation from color images that skin color models suffer from low specificity and high variance of the skin color, and this problem can be addressed by conforming the skin model to a presen
- Open Access
- 17 April 2015
- Published by Public Library of Science (PLoS) in PLOS ONE
- Vol. 10 (4) , e0121453
CoMeta: Classification of Metagenomes Using k-mers
Nowadays, the study of environmental samples has been developing rapidly. Characterization of the environment composition broadens the knowledge about the relationship between species composition and environmental conditions. An important element of extracting the knowledge of the sample composition is to compare the extracted fragments of DNA with sequences derived from known organisms. In the presented paper, we introduce an algorithm called CoMeta (Classification of metagenomes), which assigns a query read (a DNA fragment) into one of the groups previously prepared by the user. Typically, this is one of the taxonomic rank (e.g., phylum, genus), however prepared groups may contain sequences having various functions. In CoMeta, we used the exact method for read classification using short subsequences (k-mers) and fast program for indexing large set of k-mers. In contrast to the most popular methods based on BLAST, where the query is compared with each reference sequence, we begin the classification from the top of the taxonomy tree to reduce the number of comparisons. The presented experimental study confirms that CoMeta outperforms other programs used in this context. CoMeta is available at https://github.com/jkawulok/cometa under a free GNU GPL 2 license.
- 17 April 2015
- Published by Public Library of Science (PLoS) in PLOS ONE
- Vol. 10 (4) , e0121453
Open AccessCoMeta: Classification of Metagenomes Using k-mers
Nowadays, the study of environmental samples has been developing rapidly. Characterization of the environment composition broadens the knowledge about the relationship between species composition and environmental conditions. An important element of extracting the knowledge of the sample composition is to compare the extracted fragments of DNA with sequences derived from known organisms. In the presented paper, we introduce an algorithm called CoMeta (Classification of metagenomes), which assigns a query read (a DNA fragment) into one of the groups previously prepared by the user. Typically, this is one of the taxonomic rank (e.g., phylum, genus), however prepared groups may contain sequences having various functions. In CoMeta, we used the exact method for read classification using short subsequences (k-mers) and fast program for indexing large set of k-mers. In contrast to the most popular methods based on BLAST, where the query is compared with each reference sequence, we begin the classification from the top of the taxonomy tree to reduce the number of comparisons. The presented experimental study confirms that CoMeta outperforms other programs used in this context. CoMeta is available at https://github.com/jkawulok/cometa under a free GNU GPL 2 license.
- Open Access
- 29 November 2014
- Published by Springer Nature in EURASIP Journal on Advances in Signal Processing
- Vol. 2014 (1) , 170
Self-adaptive algorithm for segmenting skin regions
- 29 November 2014
- Published by Springer Nature in EURASIP Journal on Advances in Signal Processing
- Vol. 2014 (1) , 170
Open AccessSelf-adaptive algorithm for segmenting skin regions
- Open Access
- 1 May 2014
- Published by Elsevier in Pattern Recognition Letters
- Vol. 41, 3-13
Spatial-based skin detection using discriminative skin-presence features
- 1 May 2014
- Published by Elsevier in Pattern Recognition Letters
- Vol. 41, 3-13
Open AccessSpatial-based skin detection using discriminative skin-presence features