All Outputs (24)

Robust 3D U-Net Segmentation of Macular Holes (2021)
Presentation / Conference Contribution
Frawley, J., Willcocks, C. G., Habib, M., Geenen, C., Steel, D. H., & Obara, B. (2021, December). Robust 3D U-Net Segmentation of Macular Holes. Presented at The 29th Irish Conference on Artificial Intelligence and Cognitive Science 2021, Dublin, Republic of Ireland, December 9-10, 2021, Dublin, Ireland

Macular holes are a common eye condition which result in visual impairment. We look at the application of deep convolutional neural networks to the problem of macular hole segmentation. We use the 3D U-Net architecture as a basis and experiment with... Read More about Robust 3D U-Net Segmentation of Macular Holes.

Segmentation of macular edema datasets with small residual 3D U-Net architectures (2020)
Presentation / Conference Contribution
Frawley, J., Willcocks, C. G., Habib, M., Geenen, C., Steel, D. H., & Obara, B. (2020, October). Segmentation of macular edema datasets with small residual 3D U-Net architectures. Presented at 20th IEEE International Conference on BioInformatics and BioEngineering, Cincinnati, OH

This paper investigates the application of deep convolutional neural networks with prohibitively small datasets to the problem of macular edema segmentation. In particular, we investigate several different heavily regularized architectures. We find t... Read More about Segmentation of macular edema datasets with small residual 3D U-Net architectures.

Temporal neighbourhood aggregation: predicting future links in temporal graphs via recurrent variational graph convolutions (2019)
Presentation / Conference Contribution
Bonner, S., Atapour-Abarghouei, A., Jackson, P., Brennan, J., Kureshi, I., Theodoropoulos, G., McGough, S., & Obara, B. (2019, December). Temporal neighbourhood aggregation: predicting future links in temporal graphs via recurrent variational graph convolutions. Presented at IEEE International Conference on Big Data (Deep Graph Learning: Methodologies and Applications), Los Angeles, CA, USA

Graphs have become a crucial way to represent large, complex and often temporal datasets across a wide range of scientific disciplines. However, when graphs are used as input to machine learning models, this rich temporal information is frequently di... Read More about Temporal neighbourhood aggregation: predicting future links in temporal graphs via recurrent variational graph convolutions.

Coarse annotation refinement for segmentation of dot-matrix batchcodes (2019)
Presentation / Conference Contribution
Jia, N., Holder, C., Bonner, S., & Obara, B. (2019, December). Coarse annotation refinement for segmentation of dot-matrix batchcodes. Presented at IEEE International Conference on Machine Learning and Applications, Boca Raton, FL, USA

Deep Convolutional Neural Networks (CNN) have been extensively applied in various computer vision tasks. Although such approaches have demonstrated exceptionally high performance in various open challenges, adapting them to more specialised tasks can... Read More about Coarse annotation refinement for segmentation of dot-matrix batchcodes.

On the use of neural text generation for the task of optical character recognition (2019)
Presentation / Conference Contribution
Mohammadi, M., Jaf, S., Breckon, T., Matthews, P., McGough, A. S., Theodoropoulos, G., & Obara, B. (2019, November). On the use of neural text generation for the task of optical character recognition. Presented at 16th ACS/IEEE International Conference on Computer Systems and Applications AICCSA 2019., Abu Dhabi, UAE

Optical Character Recognition (OCR), is extraction of textual data from scanned text documents to facilitate their indexing, searching, editing and to reduce storage space. Although OCR systems have improved significantly in recent years, they still... Read More about On the use of neural text generation for the task of optical character recognition.

Analysing social media as a hybrid tool to detect and interpret likely radical behavioural traits for national security (2019)
Presentation / Conference Contribution
Cardenas-Canto, P., Theodoropoulos, G., Obara, B., & Kureshi, I. (2019, December). Analysing social media as a hybrid tool to detect and interpret likely radical behavioural traits for national security. Presented at IEEE International Conference on Big Data (Human-in-the-loop Methods and Human Machine Collaboration in BigData), Los Angeles, CA, USA

The study of National Security and its associated considerations is a sensitive and complex paradigm. It encapsulates both the protection of the territorial integrity and sovereignty of a state, as well as guaranteeing the security of its population.... Read More about Analysing social media as a hybrid tool to detect and interpret likely radical behavioural traits for national security.

Combining mathematical morphology and the Hilbert transform for fully automatic nuclei detection in fluorescence microscopy (2019)
Presentation / Conference Contribution
Nelson, C. J., Jackson, P. T., & Obara, B. (2019, August). Combining mathematical morphology and the Hilbert transform for fully automatic nuclei detection in fluorescence microscopy. Presented at The International Symposium on Mathematical Morphology, Saarbrucken, Germany

Phenotypic profiling of high throughput imaging screens with generic deep convolutional features (2019)
Presentation / Conference Contribution
Jackson, P., Wang, Y., Bendtsen, C., Knight, S., Chen, H., Dorval, T., Brown, M., & Obara, B. (2019, May). Phenotypic profiling of high throughput imaging screens with generic deep convolutional features. Presented at International Conference on Machine Vision Applications, Tokyo, Japan

Style Augmentation: Data Augmentation via Style Randomization (2019)
Presentation / Conference Contribution
Jackson, P., Atapour-Abarghouei, A., Bonner, S., Breckon, T., & Obara, B. (2019, June). Style Augmentation: Data Augmentation via Style Randomization. Presented at IEEE/CVF Conference on Computer Vision and Pattern Recognition, Deep Vision, Long Beach, CA, USA

We introduce style augmentation, a new form of data augmentation based on random style transfer, for improving the robustness of Convolutional Neural Networks (CNN) over both classification and regression based tasks. During training, style augmentat... Read More about Style Augmentation: Data Augmentation via Style Randomization.

TMIXT: A process flow for Transcribing MIXed handwritten and machine-printed Text (2018)
Presentation / Conference Contribution
Medhat, F., Mohammadi, M., Jaf, S., Willcocks, C., Breckon, T., Matthews, P., McGough, A. S., Theodoropoulos, G., & Obara, B. (2018, December). TMIXT: A process flow for Transcribing MIXed handwritten and machine-printed Text. Presented at IEEE International Conference on Big Data., Seattle, WA, USA

—Text recognition of scanned documents is usually dependent upon the type of text, being handwritten or machine-printed. Accordingly, the recognition involves prior classification of the text category, before deciding on the recognition method to be... Read More about TMIXT: A process flow for Transcribing MIXed handwritten and machine-printed Text.

Temporal Graph Offset Reconstruction: Towards Temporally Robust Graph Representation Learning (2018)
Presentation / Conference Contribution
Bonner, S., Brennan, J., Kureshi, I., Theodoropoulos, G., McGough, S., & Obara, B. (2018, December). Temporal Graph Offset Reconstruction: Towards Temporally Robust Graph Representation Learning. Presented at IEEE International Conference on Big Data., Seattle, WA, USA

Graphs are a commonly used construct for representing relationships between elements in complex high dimensional datasets. Many real-world phenomenon are dynamic in nature, meaning that any graph used to represent them is inherently temporal. However... Read More about Temporal Graph Offset Reconstruction: Towards Temporally Robust Graph Representation Learning.

Defining an alert mechanism for detecting likely threats to national security (2018)
Presentation / Conference Contribution
Cardenas-Canto, P., Theodoropoulos, G., & Obara, B. (2018, December). Defining an alert mechanism for detecting likely threats to national security. Presented at IEEE International Conference on Big Data., Seattle WA, USA

2D and 3D vascular structures enhancement via multiscale fractional anisotropy tensor (2018)
Presentation / Conference Contribution
Alhasson, H. F., Alharbi, S. S., & Obara, B. (2018, September). 2D and 3D vascular structures enhancement via multiscale fractional anisotropy tensor. Presented at European Conference on Computer Vision, Munich, Germany

A conceptual framework for social movements analytics for national security (2018)
Presentation / Conference Contribution
Cárdenas, P., Theodoropoulos, G., Obara, B., & Kureshi, I. (2018, June). A conceptual framework for social movements analytics for national security. Presented at International Conference on Computational Science, Wuxi, China

Social media tools have changed our world due to the way they convey information between individuals; this has led to many social movements either starting on social media or being organised and managed through this medium. At times however, certain... Read More about A conceptual framework for social movements analytics for national security.

Curvilinear structure enhancement by multiscale top-hat tensor in 2D/3D images (2018)
Presentation / Conference Contribution
Alharbi, S. S., Sazak, C., Nelson, C. J., & Obara, B. (2018, December). Curvilinear structure enhancement by multiscale top-hat tensor in 2D/3D images. Presented at IEEE International Conference on Bioinformatics and Biomedicine, Madrid, Spain

The multiscale bowler-hat transform for vessel enhancement in 3D biomedical images (2018)
Presentation / Conference Contribution
Sazak, C., Nelson, C. J., & Obara, B. (2018, September). The multiscale bowler-hat transform for vessel enhancement in 3D biomedical images. Presented at The 29th British Machine Vision Conference (BMVC 2018)., Newcastle, UK

Visual siamese clustering for cosmetic product recommendation (2018)
Presentation / Conference Contribution
Holder, C., & Obara, B. (2018, December). Visual siamese clustering for cosmetic product recommendation. Presented at 14th Asian Conference on Computer Vision (ACCV)., Perth, Australia

We investigate the problem of a visual similarity-based recommender system, where cosmetic products are recommended based on the preferences of people who share similarity of visual features. In this work we train a Siamese convolutional neural netwo... Read More about Visual siamese clustering for cosmetic product recommendation.

Evaluating the quality of graph embeddings via topological feature reconstruction (2017)
Presentation / Conference Contribution
Bonner, S., Brennan, J., Kureshi, I., Theodoropoulos, G., McGough, S., & Obara, B. (2017, December). Evaluating the quality of graph embeddings via topological feature reconstruction. Presented at 2017 IEEE International Conference on Big Data, Boston, MA

Contrast-independent curvilinear structure enhancement in 3D biomedical images (2017)
Presentation / Conference Contribution
Sazak, Ç., & Obara, B. (2017, April). Contrast-independent curvilinear structure enhancement in 3D biomedical images. Presented at IEEE International Symposium on Biomedical Imaging, Melbourne, Australia

A wide range of biomedical applications require detection, quantification and modelling of curvilinear structures in 3D images. Here we propose a 3D contrast-independent approach to enhance curvilinear structures based on the 3D Phase Congruency Tens... Read More about Contrast-independent curvilinear structure enhancement in 3D biomedical images.

Avoiding over-detection: towards combined object detection and counting (2017)
Presentation / Conference Contribution
Jackson, P. T., & Obara, B. (2017, June). Avoiding over-detection: towards combined object detection and counting. Presented at 16th International Conference on Artificial Intelligence and Soft Computing (ICAISC 2017)., Zakopane, Poland

Existing object detection frameworks in the deep learning field generally over-detect objects, and use non-maximum suppression (NMS) to filter out excess detections, leaving one bounding box per object. This works well so long as the ground-truth bou... Read More about Avoiding over-detection: towards combined object detection and counting.