Stochastic Rounding for Image Interpolation and Scan Conversion

Rukundo, Olivier; Schmidt, Samuel

doi:10.14569/ijacsa.2022.0130303

Cited by 6 publications

(7 citation statements)

References 31 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…First, it is important to remember that the nearest neighbor interpolation would be the simplest option to interpolate GT masks due to its inherent advantage of not creating non-original or extra class labels in the interpolated masks. The only problem is the deterministic rounding function on which its pixel selection strategy is based [48]. Such a strategy slightly shifts the entire image content to some extent, and is responsible for creating heavy jagged artefacts in interpolation results [35,45,48].…”

Section: A Novel Strategy For Removing Extra Class Labels In Interpol...mentioning

confidence: 99%

Effects of Image Size on Deep Learning

Rukundo

2023

Electronics

View full text Add to dashboard Cite

In this work, the best size for late gadolinium enhancement (LGE) magnetic resonance imaging (MRI) images in the training dataset was determined to optimize deep learning training outcomes. Non-extra pixel and extra pixel interpolation algorithms were used to determine the new size of the LGE-MRI images. A novel strategy was introduced to handle interpolation masks and remove extra class labels in interpolated ground truth (GT) segmentation masks. The expectation maximization, weighted intensity, a priori information (EWA) algorithm was used for the quantification of myocardial infarction (MI) in automatically segmented LGE-MRI images. Arbitrary threshold, comparison of the sums, and sums of differences are methods used to estimate the relationship between semi-automatic or manual and fully automated quantification of myocardial infarction (MI) results. The relationship between semi-automatic and fully automated quantification of MI results was found to be closer in the case of bigger LGE MRI images (55.5% closer to manual results) than in the case of smaller LGE MRI images (22.2% closer to manual results).

show abstract

Section: A Novel Strategy For Removing Extra Class Labels In Interpol...mentioning

confidence: 99%

Effects of Image Size on Deep Learning

Rukundo

2023

Electronics

View full text Add to dashboard Cite

show abstract

“…In total, ~25,000 candidate SNPs for environmental associations were kept for the analysis. Finally, we used nearest‐neighbor interpolation (as in Rukundo & Cao, 2012: custom scripts available in the GitHub repository) to predict the RONA values for the entire distribution of the two species for spatial visualization.…”

Section: Methodsmentioning

confidence: 99%

“…sis. Finally, we used nearest-neighbor interpolation (as inRukundo & Cao, 2012: custom scripts available in the GitHub repository) to predict the RONA values for the entire distribution of the two species for spatial visualization.3 | RE SULTS3.1 | Populations exhibiting a high level of admixture are spread across three climatic groupsWe used the same sampling of 55 populations covering the whole range of P. abies and the western part of the distribution range of P. obovata as inZhou et al (2023). Using a combination of population genetics analyses, the authors showed that there are two main genetic entities: one western cluster, corresponding to Norway spruce (P. abies), centered on the Alps and the Carpathians, and extending toward Scandinavia and the Russian Plains, and one eastern cluster stemming from Siberia, east of the Urals Mountains and corresponding to what has generally been recognized as Siberian spruce (P. obovata).…”

mentioning

confidence: 99%

Hybridization mediated range expansion and climate change resilience in two keystone tree species of boreal forests

Karunarathne,

Zhou,

Lascoux

et al. 2024

Global Change Biology

View full text Add to dashboard Cite

Current global climate change is expected to affect biodiversity negatively at all scales leading to mass biodiversity loss. Many studies have shown that the distribution of allele frequencies across a species' range is often influenced by specific genetic loci associated with local environmental variables. This association reflects local adaptation and allele changes at those loci could thereby contribute to the evolutionary response to climate change. However, predicting how species will adapt to climate change from this type of data alone remains challenging. In the present study, we combined exome capture sequences and environmental niche reconstruction, to test multiple methods for assessing local adaptation and climate resilience in two widely distributed conifers, Norway spruce and Siberian spruce. Both species are keystone species of the boreal forest and share a vast hybrid zone. We show that local adaptation in conifers can be detected through allele frequency variation, population‐level ecological preferences, and historical niche movement. Moreover, we integrated genetic and ecological information into genetic offset predictive models to show that hybridization plays a central role in expanding the niche breadth of the two conifer species and may help both species to cope better with future changing climates. This joint genetic and ecological analysis also identified spruce populations that are at risk under current climate change.

show abstract

“…As shown in the Figure 2c, decrease in the width of the bar pattern is visible which is due to finite length of the tracks with an increase in the track density. The simulated images were scaled down to 188 pixels × 188 pixels using the nearest neighbor method 46 where size of one pixel is equivalent to 0.3 µm (typical diameter of a silver grain). Figures 2d, 2e and 2f show resized and inverted images of Figures 2a, 2b and 2c, respectively.…”

Section: A Estimation Of Track Densitymentioning

confidence: 99%

Investigation of the neutron imaging applications using fine-grained nuclear emulsion

Muneem¹,

Yoshida²,

Ekawa³

et al. 2022

Preprint

View full text Add to dashboard Cite

Neutron imaging is a non-destructive inspection technique with a wide range of applications. One of the important aspects concerning neutron imaging is achieving micrometer-scale spatial resolution. Developing a neutron detector with a high resolution is a challenging task. Neutron detectors, based on fine-grained nuclear emulsion, may be suitable for high resolution neutron imaging applications. High track density is a necessary requirement to improve the quality of neutron imaging. However, the available track analysis methods are difficult to apply under high track density conditions. Simulated images were used to determine the required track density for neutron imaging. It was concluded that a track density of the order of 10 4 tracks per 100 × 100 µm 2 is sufficient to utilize neutron detectors for imaging applications. The contrast resolution was also investigated for the image data sets with various track densities and neutron transmission rates. Moreover, experiments were performed for neutron imaging of the gadolinium-based gratings with known geometries. The structure of gratings was successfully resolved. The calculated 1σ 10-90 % edge response, using the gray scale optical images of the grating slit with a periodic structure of 9 µm, was 0.945 ± 0.004 µm.

show abstract

Stochastic Rounding for Image Interpolation and Scan Conversion

Cited by 6 publications

References 31 publications

Effects of Image Size on Deep Learning

Effects of Image Size on Deep Learning

Hybridization mediated range expansion and climate change resilience in two keystone tree species of boreal forests

Investigation of the neutron imaging applications using fine-grained nuclear emulsion

Contact Info

Product

Resources

About