Michał Romaszewski

Dr. Michał Romaszewski

orcid

Research Focus & Metrics

Academic Career Length

18Years (2009-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
1 November 2016 in ISPRS Journal of Photogrammetry and Remote Sensing
Verified
Journal Article
Closed access
MRMichał Romaszewski
Michał Romaszewski
PGPrzemysław Głomb
Przemysław Głomb
MCMichał Cholewa
Michał Cholewa

1 September 2018 in Forensic Science International
Verified
Journal Article
Closed access
PGPrzemysław Głomb
Przemysław Głomb
MRMichał Romaszewski
Michał Romaszewski
MCMichał Cholewa
Michał Cholewa
Krzysztof A. DOMINOKrzysztof A. DOMINO
Krzysztof A. DOMINO

21 November 2020 in Sensors
Verified
Journal Article
Open access
PDF file
KKKamil Książek
Kamil Książek
MRMichał Romaszewski
Michał Romaszewski
PGPrzemysław Głomb
Przemysław Głomb
BGBartosz Grabowski
Bartosz Grabowski
MCMichał Cholewa
Michał Cholewa
In recent years, growing interest in deep learning neural networks has raised a question on how they can be used for effective processing of high-dimensional datasets produced by hyperspectral imaging (HSI). HSI, traditionally viewed as being within the scope of remote sensing, is used in non-invasive substance classification. One of the areas of potential application is forensic science, where substance classification on the scenes is important. An example problem from that area—blood stain classification—is a case study for the evaluation of methods that process hyperspectral data. To investigate the deep learning classification performance for this problem we have performed experiments on a dataset which has not been previously tested using this kind of model. This dataset consists of several images with blood and blood-like substances like ketchup, tomato concentrate, artificial blood, etc. To test both the classic approach to hyperspectral classification and a more realistic application-oriented scenario, we have prepared two different sets of experiments. In the first one, Hyperspectral Transductive Classification (HTC), both a training and a test set come from the same image. In the second one, Hyperspectral Inductive Classification (HIC), a test set is derived from a different image, which is more challenging for classifiers but more useful from the point of view of forensic investigators. We conducted the study using several architectures like 1D, 2D and 3D convolutional neural networks (CNN), a recurrent neural network (RNN) and a multilayer perceptron (MLP). The performance of the models was compared with baseline results of Support Vector Machine (SVM). We have also presented a model evaluation method based on t-SNE and confusion matrix analysis that allows us to detect and eliminate some cases of model undertraining. Our results show that in the transductive case, all models, including the MLP and the SVM, have comparative performance, with no clear advantage of deep learning models. The Overall Accuracy range across all models is 98–100% for the easier image set, and 74–94% for the more difficult one. However, in a more challenging inductive case, selected deep learning architectures offer a significant advantage; their best Overall Accuracy is in the range of 57–71%, improving the baseline set by the non-deep models by up to 9 percentage points. We have presented a detailed analysis of results and a discussion, including a summary of conclusions for each tested architecture. An analysis of per-class errors shows that the score for each class is highly model-dependent. Considering this and the fact that the best performing models come from two different architecture families (3D CNN and RNN), our results suggest that tailoring the deep neural network architecture to hyperspectral data is still an open problem.

1 March 2021 in Forensic Science International
Verified
Journal Article
Closed access
MRMichał Romaszewski
Michał Romaszewski
PGPrzemysław Głomb
Przemysław Głomb
ASArkadiusz Sochan
Arkadiusz Sochan
MCMichał Cholewa
Michał Cholewa

24 November 2020 in Animals
Verified
Journal Article
Open access
PDF file
MDMalgorzata Domino
Malgorzata Domino
MRMichał Romaszewski
Michał Romaszewski
TJTomasz Jasiński
Tomasz Jasiński
MMMałgorzata Maśko
Małgorzata Maśko
Infrared thermography (IRT) is a valuable diagnostic tool in equine veterinary medicine; however, little is known about its application to donkeys. This study aims to find patterns in thermal images of donkeys and horses and determine if these patterns share similarities. The study is carried out on 18 donkeys and 16 horses. All equids undergo thermal imaging with an infrared camera and measurement of the skin thickness and hair coat length. On the class maps of each thermal image, fifteen regions of interest (ROIs) are annotated and then combined into 10 groups of ROIs (GORs). The existence of statistically significant differences between surface temperatures in GORs is tested both “globally” for all animals of a given species and “locally” for each animal. Two special cases of animals that differed from the rest are also discussed. The results indicate that the majority of thermal patterns are similar for both species; however, average surface temperatures in horses (22.72±2.46 °C) are higher than in donkeys (18.88±2.30 °C). This could be related to differences in the skin thickness and hair coat. The patterns of both species are associated with GORs, rather than with an individual ROI, and there is a higher uniformity in the donkeys’ patterns.

2 June 2025 in Bulletin Of The Polish Academy Of Sciences-Technical Sciences
Verified
Journal Article
Open access
MGManish K. Gupta
Manish K. Gupta
MRMichał Romaszewski
Michał Romaszewski
PGPiotr Gawron
Piotr Gawron
Quantum computers with hundreds of noisy qubits are already available for the research community. They have the potential to run complex quantum computations well beyond the computational capacity of any classical device. It is natural to ask the question, what application these devices could be useful for? Land use and land cover classification of multispectral Earth observation data collected from the earth observation satellite mission is one such problem that is hard for classical methods due to its unique characteristics. In this work, we compare the performance of several classical machine learning algorithms on the stilted re-labeled dataset of the Copernicus Sentinel-2 mission, when the algorithm has access to projected quantum kernel (PQK) features. We show that the classification accuracy increases drastically when the model has access to PQK features. We then naively study the performance of these algorithms with and without access to PQK features on the original Copernicus Sentinel-2 mission data set. This study provides key evidence that shows the potential of quantum machine learning methods for Earth observation data.