Browsing by Subject "Convolutional neural networks"

Now showing 1 - 14 of 14

Open Access
Channel estimation and symbol demodulation for OFDM systems over rapidly varying multipath channels with hybrid deep neural networks
(Institute of Electrical and Electronics Engineers, 2023-05-01) Gümüş, Mücahit; Duman, Tolga Mete
We consider orthogonal frequency division multiplexing over rapidly time-varying multipath channels, for which performance of standard channel estimation and equalization techniques degrades dramatically due to inter-carrier interference (ICI). We focus on improving the overall system performance by designing deep neural network (DNN) architectures for both channel estimation and data demodulation. To accomplish this, we employ the basis expansion model to track the channel tap variations, and exploit convolutional neural networks’ learning abilities of local correlations together with a coarse least square solution for a robust and accurate channel estimation procedure. For data demodulation, we use a recurrent neural network for improved performance and robustness as single tap frequency-domain equalizers perform poorly, and more sophisticated equalization techniques such as band-limited linear minimum mean squared error equalizers are vulnerable to model mismatch and channel estimation errors. Numerical examples illustrate that the proposed DNN architectures outperform the traditional algorithms. Specifically, the bit error rate results for a wide range of Doppler values reveal that the proposed DNN-based equalizer is robust, and it mitigates the ICI effectively, offering an excellent demodulation performance. We further note that the DNN-based channel estimator offers an improved performance with a reduced computational complexity.
Open Access
A complete framework of radar pulse detection and modulation classification for cognitive EW
(Institute of Electrical and Electronics Engineers Inc., 2019) Yar, Ersin; Kocamış, M. B.; Orduyılmaz, A.; Serin, M.; Efe, M.
In this paper, we consider automatic radar pulse detection and intra-pulse modulation classification for cognitive electronic warfare applications. In this manner, we introduce an end-to-end framework for detection and classification of radar pulses. Our approach is complete, i.e., we provide raw radar signal at the input side and produce categorical output at the output. We use short time Fourier transform to obtain time-frequency image of the signal. Hough transform is used to detect pulses in time-frequency images and pulses are represented with a single line. Then, convolutional neural networks are used for pulse classification. In experiments, we provide classification results at different SNR levels.
Open Access
Deep clustering via center-oriented margin free-triplet loss for skin lesion detection in highly ımbalanced datasets
(Institute of Electrical and Electronics Engineers Inc., 2022-06-29) Öztürk, Şaban; Çukur, Tolga
Melanoma is a fatal skin cancer that is curable and has dramatically increasing survival rate when diagnosed at early stages. Learning-based methods hold significant promise for the detection of melanoma from dermoscopic images. However, since melanoma is a rare disease, existing databases of skin lesions predominantly contain highly imbalanced numbers of benign versus malignant samples. In turn, this imbalance introduces substantial bias in classification models due to the statistical dominance of the majority class. To address this issue, we introduce a deep clustering approach based on the latent-space embedding of dermoscopic images. Clustering is achieved using a novel center-oriented margin-free triplet loss (COM-Triplet) enforced on image embeddings from a convolutional neural network backbone. The proposed method aims to form maximally-separated cluster centers as opposed to minimizing classification error, so it is less sensitive to class imbalance. To avoid the need for labeled data, we further propose to implement COM-Triplet based on pseudo-labels generated by a Gaussian mixture model (GMM). Comprehensive experiments show that deep clustering with COM-Triplet loss outperforms clustering with triplet loss, and competing classifiers in both supervised and unsupervised settings. © 2013 IEEE.
Open Access
Deep fractional Fourier networks
(2024-08) Koç, Emirhan
This thesis introduces the integration of the fractional Fourier Transform (FrFT) into the deep learning domain, with the aim of opening new avenues for incorporating signal processing into deep neural networks (DNNs). This work starts by introducing FrFT into recurrent neural networks (RNNs) for time series prediction, leveraging its ability and flexibility to perform infinitely many continuous transformations and offering an alternative to the traditional Fourier Transform (FT). Despite the initial success, a significant challenge identified is the manual tuning of the fraction order parameter a, which can be cumbersome and limits broader applicability. To overcome this limitation, we introduce a novel approach where the fraction order a is treated as a learnable parameter within deep learning models. First, a theoretical foundation is established to support the learnability of this parameter, followed by extensive experimentation in image classification and time series prediction tasks. The results demonstrate that incorporating a learnable fraction order significantly improves model performance, particularly when integrated with well-known architectures such as ResNet and VGG models. Furthermore, the thesis proposes fractional Fourier Pooling (FrFP), a pooling technique that replaces traditional Global Average Pooling (GAP) layers in Convolutional Neural Networks (CNNs). FrFP enhances feature representation by processing intermediate signal regions, leading to better model performance and offering a new perspective on integrating signal transformations within deep learning frameworks. Overall, this thesis contributes to the growing body of research exploring advanced signal processing techniques in deep learning, highlighting the potential of FrFT as a powerful tool for improving model accuracy and efficiency across various applications.
Open Access
Deep learning based unsupervised tissue segmentation in histopathological images
(2017-11) Köylü, Troya Çağıl
In the current practice of medicine, histopathological examination of tissues is essential for cancer diagnosis. However, this task is both subject to observer variability and time consuming for pathologists. Thus, it is important to develop automated objective tools, the first step of which usually comprises image segmentation. According to this need, in this thesis, we propose a novel approach for the segmentation of histopathological tissue images. Our proposed method, called deepSeg, is a two-tier method. The first tier transfers the knowledge from AlexNet, which is a convolutional neural network (CNN) trained for the non-medical domain of ImageNet, to the medical domain of histopathological tissue image characterization. The second tier uses this characterization in a seed-controlled region growing algorithm, for the unsupervised segmentation of heterogeneous tissue images into their homogeneous regions. To test the effectiveness of the segmentation, we conduct experiments on microscopic colon tissue images. Quantitative results reveal that the proposed method improves the performance of the previous methods that work on the same dataset. This study both illustrates one of the first successful demonstrations of using deep learning for tissue image segmentation, and shows the power of using deep learning features instead of handcrafted ones in the domain of histopathological image analysis.
Open Access
Deep learning in electronic warfare systems: Automatic intra-pulse modulation recognition
(Institute of Electrical and Electronics Engineers, 2018) Akyön, Fatih Çağatay; Alp, Y. K.; Gök, G.; Arıkan, Orhan
Detection and classification of radars in electronic warfare systems is a major problem. In this work, we propose a novel deep learning based method that automatically recognizes intra-pulse modulation types of radar signals. We use reassigned short-time Fourier transforms of measured signals and detected outliers of the phase differences using robust least squares to train a hybrid structured convolutional neural network to distinguish frequency and phase modulated signals. Simulation results show that the developed method highly outperforms the current state-of-the-art methods in the literature.
Open Access
Detection of jammers in range-doppler images generated in DTED based radar simulator using convolutional neural networks
(IEEE - Institute of Electrical and Electronics Engineers, 2023-08-28) Şahinbay, H. E.; Akyol, Ali Alp; Özdemir, Ö.
Airborne radars have a variety of air-to-air and air-to-ground missions. In both air-to-air and air-to-ground target detection missions, ground clutter reflections are received from the main beam and side lobes of the radar. The effects of this clutter can be clearly seen in the radar range-Doppler maps. In addition, there may be other sources in the environment that distort the radar's range-Doppler maps. These sources can be categorized as jammer and interference signals. They distord the range-Doppler maps, making target detection more difficult, interfering with target detection and, in some cases, leading to false target detection. The detection of jammer and interference signals, which are the source of this situation, is of critical importance for the operators controlling the platform. It is often not possible for operators to quickly detect and classify these jamming signals. Deep learning methods, which have recently been used in every field, can achieve much faster and robust target detection and classification results compared to humans. In this study, the success of a Convolutional Neural Network based technique, which is one of the deep learning methods, in detecting and classifying jammer and interference signals is investigated.
Open Access
Image classification with energy efficient hadamard neural networks
(2018-01) Deveci, Tuba Ceren
Deep learning has made significant improvements at many image processing tasks in recent years, such as image classification, object recognition and object detection. Convolutional neural networks (CNN), which is a popular deep learning architecture designed to process data in multiple array form, show great success to almost all detection & recognition problems and computer vision tasks. However, the number of parameters in a CNN is too high such that the computers require more energy and larger memory size. In order to solve this problem, we investigate the energy efficient network models based on CNN architecture. In addition to previously studied energy efficient models such as Binary Weight Network (BWN), we introduce novel energy efficient models. Hadamard-transformed Image Network (HIN) is a variation of BWN, but uses compressed Hadamardtransformed images as input. Binary Weight and Hadamard-transformed Image Network (BWHIN) is developed by combining BWN and HIN as a new energy ef- ficient model. Performances of the neural networks with di erent parameters and di erent CNN architectures are compared and analyzed on MNIST and CIFAR-10 datasets. It is observed that energy efficiency is achieved with a slight sacrifice at classification accuracy. Among all energy efficient networks, our novel ensemble model outperforms other energy efficient models.
Open Access
Manyetik parçacık görüntüleme için evrişimsel sinir ağı tabanlı bir süper-çözünürlük tekniği
(IEEE, 2021-07-19) Aşkın, Barış; Güngör, Alper; Soydan, Damla Alptekin; Top, Can Barış; Çukur, Tolga
Manyetik Parçacık Görüntüleme (MPG), süperparamanyetik demir-oksit (SPDO) parçacıklarının yüksek çözünürlük ve kare hızında görüntülenmesini sağlayan bir görüntüleme yöntemidir. Görüntüleme işlemi doğrusal olarak modellenebilmektedir. Ancak deneysel sistemlerin ideal dışı davranışı ve teorik sistemlere kıyasla değişimlerinden dolayı, MPG sistemlerinde çoğu durumda öncelikli olarak ileri model matrisi ölçülür (sistem kalibre edilir) ve ardından bu matrisler kullanılarak görüntülerin geriçatımı yapılır. Görüntü çözünürlüğü ve boyutu doğrudan sistem matrisinin boyutundan etkilenmektedir. Ancak, kalibrasyon işlemi görüntüleme alanına bağlı olarak çok zaman almaktadır. Bu çalışmada, düşük çözünürlükte ölçülen sistem matrisleri üzerinde süper-çözünürlük teknikleri kullanılarak yüksek çözünürlüklü sistem matrisi elde edilmesi önerilmektedir. Bu amaç doğrultusunda evrişimsel sinir ağı (ESA) tabanlı bir süperçözünürlük tekniği MPG için uyarlanmış ve doğrusal aradeğerlemeye (interpolasyon) karşı etkinliği gösterilmiştir. Yöntemler gürültüsüz bir benzetim ortamında kıyaslanmış ve 4 4 kat süper-çözünürlük için, önerilen yöntem %2.92 normalize edilmiş ortalama kare hatasına yol açarken, bikübik aradeğerlemenin %12.47 hataya yol açtığı gösterilmiştir.
Open Access
Segmentation-aware MRI reconstruction
(Springer Cham, 2022-09-22) Acar, Mert; Çukur, Tolga; Öksüz, İ.
Deep learning models have been broadly adopted for accelerating MRI acquisitions in recent years. A common approach is to train deep models based on loss functions that place equal emphasis on reconstruction errors across the field-of-view. This homogeneous weighting of loss contributions might be undesirable in cases where the diagnostic focus is on tissues in a specific subregion of the image. In this paper, we propose a framework for segmentation-aware reconstruction based on segmentation as a proxy task. We leverage an end-to-end model comprising reconstruction and segmentation networks; and leverage backpropagation of segmentation error to devise a pseudo-attention effect to focus the reconstruction network. We introduce a novel stabilization method to prevent convergence onto a local minima with unacceptably poor reconstruction or segmentation performance. Our stabilization approach initiates learning on fully-sampled acquisitions, and gradually increases the undersampling rate assumed in the training set to its desired value. We validate our approach for cardiac MR reconstruction on the publicly available OCMR dataset. Segmentation-aware reconstruction significantly outperforms vanilla reconstruction for cardiac imaging.
Open Access
Self-supervised dynamic MRI reconstruction
(Springer, 2021-09-25) Acar, Mert; Çukur, Tolga; Öksüz, İlkay
Deep learning techniques have recently been adopted for accelerating dynamic MRI acquisitions. Yet, common frameworks for model training rely on availability of large sets of fully-sampled MRI data to construct a ground-truth for the network output. This heavy reliance is undesirable as it is challenging to collect such large datasets in many applications, and even impossible for high spatiotemporal-resolution protocols. In this paper, we introduce self-supervised training to deep neural architectures for dynamic reconstruction of cardiac MRI. We hypothesize that, in the absence of ground-truth data, elevating complexity in self-supervised models can instead constrain model performance due to the deficiencies in training data. To test this working hypothesis, we adopt self-supervised learning on recent state-of-the-art deep models for dynamic MRI, with varying degrees of model complexity. Comparison of supervised and self-supervised variants of deep reconstruction models reveals that compact models have a remarkable advantage in reliability against performance loss in self-supervised settings.
Open Access
Shape-preserving loss in deep learning for cell segmentation
(2020-07) Hüseyin, Furkan
Fully convolutional networks (FCNs) have become the state-of-the-art models for cell instance segmentation in microscopy images. These networks are trained by minimizing a loss function, which typically deﬁnes the loss of each pixel separately and aggregates these pixel losses by averaging or summing. Since this pixel-wise deﬁnition of a loss function does not consider the spatial relations between the pixels’ predictions, it does not suﬃciently impose the network to learn a particular shape(s). On the other hand, this ability of the network might be important for better segmenting cells, which commonly show similar morphological characteristics due to their natures. In response to this issue, this thesis introduces a new dynamic shape-preserving loss function to train an FCN for cell instance segmentation. This loss function is a weighted cross-entropy whose pixel weights are deﬁned as prior-shape-aware. To this end, it calculates the weights based on the similarity between the shape of the segmented objects that the pixels belong to and the shape-priors estimated on the ground truth cells. This thesis uses Fourier descriptors to quantify the shape of a cell and proposes to deﬁne a similarity metric on the distribution of these Fourier descriptors. Working on four diﬀerent medical image datasets, the experimental results demonstrate that this proposed loss function outperforms its counterpart for the segmentation of instances in these datasets.
Open Access
Spatio-temporal forecasting over graphs with deep learning
(2020-12) Ceyani, Emir
We study spatiotemporal forecasting of high-dimensional rectangular grid graph structured data, which exhibits both complex spatial and temporal dependencies. In most high-dimensional spatiotemporal forecasting scenarios, deep learningbased methods are widely used. However, deep learning algorithms are overconfident in their predictions, and this overconfidence causes problems in the human-in-the-loop domains such as medical diagnosis and many applications of 5 th generation wireless networks. We propose spatiotemporal extensions to variational autoencoders for regularization, robustness against out-of data distribution, and incorporating uncertainty in predictions to resolve overconfident predictions. However, variational inference methods are prone to biased posterior approximations due to using explicit exponential family densities and mean-field assumption in their posterior factorizations. To mitigate these problems, we utilize variational inference & learning with semi-implicit distributions and apply this inference scheme into convolutional long-short term memory networks(ConvLSTM) for the first time in the literature. In chapter 3, we propose variational autoencoders with convolutional long-short term memory networks, called VarConvLSTM. In chapter 4, we improve our algorithm via semi-implicit & doubly semi-implicit variational inference to model multi-modalities in the data distribution . In chapter 5, we demonstrate that proposed algorithms are applicable for spatiotemporal forecasting tasks, including space-time mobile traffic forecasting over Turkcell base station networks.
Open Access
Spoofing attack detection by anomaly detection
(Institute of Electrical and Electronics Engineers Inc., 2019) Fatemifar, S.; Arashloo, Shervin Rahimzadeh; Awais, M.; Kittler, J.
Spoofing attacks on biometric systems can seriously compromise their practical utility. In this paper we focus on face spoofing detection. The majority of papers on spoofing attack detection formulate the problem as a two or multiclass learning task, attempting to separate normal accesses from samples of different types of spoofing attacks. In this paper we adopt the anomaly detection approach proposed in [1], where the detector is trained on genuine accesses only using one-class classifiers and investigate the merit of subject specific solutions. We show experimentally that subject specific models are superior to the commonly used client independent method. We also demonstrate that the proposed approach is more robust than multiclass formulations to unseen attacks.