Yazar "Uslu, Fatmatulzehra" seçeneğine göre listele

Listeleniyor 1 - 6 / 6
  • Küçük Resim Yok
    Öğe
    A Comparative Study for 6D Pose Estimation of Textureless and Symmetric Objects Used in Automotive Manufacturing Industry
    (Institute of Electrical and Electronics Engineers Inc., 2023) Doruk, Abdullah Enes; Ozkaya, Tayyip Ensar; Gulmez, Furkan; Uslu, Fatmatulzehra
    6D pose estimation of industrial objects on RGB images has a high potential to accelerate the automation of robotic manipulations in the automotive manufacturing industry. Despite its high potential, this problem has not been adequately addressed in the computer vision community. Main factors leading to under investigation of this problem are industrial objects to be textureless, thin, and symmetrical, which hinder the automatic estimation of their poses from color images. Deep learning models have shown promising results for pose estimation of household objects thanks to availability of large datasets with labels. In contrast to many household objects, there are few datasets for industrial objects with limited representation capacity, which restricts the use of deep models in pose estimation of industrial objects. In this study, we examine the eligibility of deep models on 6D pose estimation of industrial objects used in the automotive manufacturing industry. For this aim, we compare the performance of three deep models, DeepIM, CosyPose, and EPOS. To meet the need for large training dataset of these models, we produce a large synthetic dataset from the CAD data of the industrial objects. We also collect a small real dataset for training and performance evaluation purposes. We find that CosyPose outperforms other methods with a large margin, by showing its potential to solve such a hard problem. We also observe that training models with both synthetic and real images yield the best results. © 2023 IEEE.
  • Küçük Resim Yok
    Öğe
    A robust quality estimation method for medical image segmentation with small datasets
    (Elsevier Sci Ltd, 2024) Uslu, Fatmatulzehra; Varela, Marta
    The quality check of automatic image analysis results is a necessity to eliminate poor outcomes. There are few work on segmentation quality estimation when there is no reference mask available. Generative -based and regression approaches generally have strict assumptions on shape/contrast of anatomical structures, which may fail when there are abnormalities in images or in the presence of domain shift. Ensemble approach promises more generalisability to various types of images; however, they are costly to train. Also, none of these methods were designed for small datasets. To address these shortcomings, this paper presents a segmentation quality estimation method for small datasets with arbitrary dimensions, which is validated with 2 D, 3 D and 4 D image datasets with roughly 20 training images, and on the segmentation of retinal vessel and left atrium segmentation. Although our method uses an ensemble for image segmentation, its design reduces its parameter size. We describe possible scenarios relating the amount of agreement across the base models' outputs in an ensemble to quality scores; then, present a technique to deal with high quality score estimation for poor segmentation as a result of the base models to largely agree on mistakes. We assess the performance of our method in the presence of different sources of domain shift, and compare it with methods selected from the aforementioned approaches. We found robust quality score estimation, generalisable to different datasets. Our code would be available upon acceptance.
  • Küçük Resim Yok
    Öğe
    An Evidential Mask Transformer for Left Atrium Segmentation
    (2024) Uslu, Fatmatulzehra
    The segmentation of the left atrium (LA) is required to calculate the clinical parameters of the LA, to identify diseases related to its remodeling. Generally, convolutional networks have been used for this task. However, their performance may be limited as a result of the use of local convolution operations for feature extraction. Also, such models usually need extra steps to provide uncertainty maps such as multiple forward passes for Monte Carlo dropouts or training multiple models for ensemble learning. To address these issues, we adapt mask transformers for LA segmentation which effectively use both local and global information, and train them with evidential learning to generate uncertainty maps from the learned Dirichlet distribution, with a single forward pass. We validated our approach on the STACOM 2013 dataset and found that our method can produce better segmentation performance than baseline models, and can identify locations our model’s responses are not trustable.
  • Küçük Resim Yok
    Öğe
    Detecting Aortic Valve Pathology from the 3-Chamber Cine Cardiac MRI View
    (Springer International Publishing Ag, 2022) Vimalesvaran, Kavitha; Uslu, Fatmatulzehra; Zaman, Sameer; Galazis, Christoforos; Howard, James; Cole, Graham; Bharath, Anil A.
    Cardiac magnetic resonance (CMR) is the gold standard for quantification of cardiac volumes, function, and blood flow. Tailored MR pulse sequences define the contrast mechanisms, acquisition geometry and timing which can be applied during CMR to achieve unique tissue characterisation. It is impractical for each patient to have every possible acquisition option. We target the aortic valve in the three-chamber (3-CH) cine CMR view. Two major types of anomalies are possible in the aortic valve. Stenosis: the narrowing of the valve which prevents an adequate outflow of blood, and insufficiency (regurgitation): the inability to stop the back-flow of blood into the left ventricle. We develop and evaluate a deep learning system to accurately classify aortic valve abnormalities to enable further directed imaging for patients who require it. Inspired by low level image processing tasks, we propose a multi-level network that generates heat maps to locate the aortic valve leaflets' hinge points and aortic stenosis or regurgitation jets. We trained and evaluated all our models on a dataset of clinical CMR studies obtained from three NHS hospitals (n = 1,017 patients). Our results (mean accuracy = 0.93 and F1 score = 0.91), show that an expert-guided deep learning-based feature extraction and a classification model provide a feasible strategy for prescribing further, directed imaging, thus improving the efficiency and utility of CMR scanning.
  • Küçük Resim Yok
    Öğe
    GSM-Net: A global sequence modelling network for the segmentation of short axis CINE MRI images
    (Pergamon-Elsevier Science Ltd, 2023) Uslu, Fatmatulzehra
    Atrial Fibrillation (AF) is a disease where the atria fail to properly contract but quiver instead, due to the abnormal electrical activity of the atrial tissue. In AF patients, anatomical and functional parameters of the left atrium (LA) largely differ from that of healthy people due to LA remodelling, which can continue in many cases after the catheter ablation treatment. Therefore, it is important to follow up with AF patients to detect any recurrence. LA segmentation masks obtained from short-axis CINE MRI images are used as the gold standard for the quantification of LA parameters. Thick slices of CINE MRI images hinder the use of 3D networks for segmentation while 2D architectures often fail to model inter-slice dependencies. This study presents GSM-Net which approximates 3D networks with effective modelling of inter-slice similarities with two new modules: global slice sequence encoder (GSSE) and sequence dependent channel attention module (SdCAt). In contrast to previous work modelling only local inter-slice similarities, GSSE also models global spatial dependencies across slices. SdCAt generates a distribution of attention weights over MRI slices per channel, to better trace characteristic changes in the size of the LA or other structures across slices. We found that GSM-Net outperforms previous methods on LA segmentation and helps to identify AF recurrence patients. We believe that GSM-Net can be used as an automatic tool to estimate LA parameters such as ejection fraction to identify AF, and to follow up with patients after treatment to detect any recurrence.
  • Küçük Resim Yok
    Öğe
    Saw-YOLOv5: Scale-Aware YOLOv5 for Object Detection in Aerial Images
    (Institute of Electrical and Electronics Engineers Inc., 2023) Doruk, Abdullah Enes; Algul, Mucteba; Akyurek, Feyzullah; Alpaydm, Osman Kursat; Uslu, Fatmatulzehra
    The detection of objects in aerial images is impor-tant for many real world problems related to military defense, transportation, and etc. However, this is a challenging task as a result of the presence of various scales of objects in the same image, the large variety of contexts across aerial images, various brightness levels due to image acquisition at different times of the day and so on. To address these challenges, this paper introduces Saw-YOLOv5 for object detection in aerial images. Saw-YOLOv5 is a deep network based on YOLOv5, which was proposed for object detection in natural images. Saw-YOLOv5 extends YOLOv5 with the addition of several attention modules in its design. The results of our experiments, conducted on the aerial dataset delivered by the Turkey Technology Team for the Artificial Intelligence in Transportation Competition, showed that Saw-YOLOv5 outperforms previous methods, particularly for pedestrian detection, by yielding a mean mAP of 80.23% over all objects. © 2023 IEEE.