Structured Illumination Microscopy is a widespread methodology to image live and fixed biological structures smaller than the diffraction limits of conventional optical microscopy. Using recent advances in image up-scaling through deep learning models, we demonstrate a method to reconstruct 3D SIM image stacks with twice the axial resolution attainable through conventional SIM reconstructions. We further evaluate our method for robustness to noise & generalisability to varying observed specimens, and discuss potential adaptions of the method to further improvements in resolution.
Ultrasonic imaging algorithms used in many clinical and industrial applications consist of three steps: A data pre-processing, an image formation and an image post-processing step. For efficiency, image formation often relies on an approximation of the underlying wave physics. A prominent example is the Delay-And-Sum (DAS) algorithm used in reflectivity-based ultrasonic imaging. Recently, deep neural networks (DNNs) are being used for the data pre-processing and the image post-processing steps separately. In this work, we propose a novel deep learning architecture that integrates all three steps to enable end-to-end training. We examine turning the DAS image formation method into a network layer that connects data pre-processing layers with image post-processing layers that perform segmentation. We demonstrate that this integrated approach clearly outperforms sequential approaches that are trained separately. While network training and evaluation is performed only on simulated data, we also showcase the potential of our approach on real data from a non-destructive testing scenario.
An unsupervised point cloud registration method, called salient points analysis (SPA), is proposed in this work. The proposed SPA method can register two point clouds effectively using only a small subset of salient points. It first applies the PointHop++ method to point clouds, finds corresponding salient points in two point clouds based on the local surface characteristics of points and performs registration by matching the corresponding salient points. The SPA method offers several advantages over the recent deep learning based solutions for registration. Deep learning methods such as PointNetLK and DCP train end-to-end networks and rely on full supervision (namely, ground truth transformation matrix and class label). In contrast, the SPA is completely unsupervised. Furthermore, SPA's training time and model size are much less. The effectiveness of the SPA method is demonstrated by experiments on seen and unseen classes and noisy point clouds from the ModelNet-40 dataset.
Deep learning techniques have considerably improved speech processing in recent years. Speaker representations extracted by deep learning models are being used in a wide range of tasks such as speaker recognition and speech emotion recognition. Attention mechanisms have started to play an important role in improving deep learning models in the field of speech processing. Nonetheless, despite the fact that important speaker-related information can be embedded in individual frequency-bins of the input spectral representations, current attention models are unable to attend to fine-grained information items in spectral representations. In this paper we propose Fine-grained Early Frequency Attention (FEFA) for speaker representation learning. Our model is a simple and lightweight model that can be integrated into various CNN pipelines and is capable of focusing on information items as small as frequency-bins. We evaluate the proposed model on three tasks of speaker recognition, speech emotion recognition, and spoken digit recognition. We use Three widely used public datasets, namely VoxCeleb, IEMOCAP, and Free Spoken Digit Dataset for our experiments. We attach FEFA to several prominent deep learning models and evaluate its impact on the final performance. We also compare our work with other related works in the area. Our experiments show that by adding FEFA to different CNN architectures, performance is consistently improved by substantial margins, and the models equipped with FEFA outperform all the other attentive models. We also test our model against different levels of added noise showing improvements in robustness and less sensitivity compared to the backbone networks.
Timely cancer reporting data are required in order to understand the impact of cancer, inform public health resource planning and implement cancer policy especially in Sub Saharan Africa where the reporting lag is behind world averages. Unstructured pathology reports, which contain tumor specific data, are the main source of information collected by cancer registries. Due to manual processing and labelling of pathology reports using the International Classification of Disease for oncology (ICD-O) codes, by human coders employed by cancer registries, has led to a considerable lag in cancer reporting. We present a hierarchical deep learning classification method that employs convolutional neural network models to automate the classification of 1813 anonymized breast cancer pathology reports with applicable ICD-O morphology codes across 9 classes. We demonstrate that the hierarchical deep learning classification method improves on performance in comparison to a flat multiclass CNN model for ICD-O morphology classification of the same reports.
Today's deep learning models are primarily trained on CPUs and GPUs. Although these models tend to have low error, they consume high power and utilize large amount of memory owing to double precision floating point learning parameters. Beyond the Moore's law, a significant portion of deep learning tasks would run on edge computing systems, which will form an indispensable part of the entire computation fabric. Subsequently, training deep learning models for such systems will have to be tailored and adopted to generate models that have the following desirable characteristics: low error, low memory, and low power. We believe that deep neural networks (DNNs), where learning parameters are constrained to have a set of finite discrete values, running on neuromorphic computing systems would be instrumental for intelligent edge computing systems having these desirable characteristics. To this extent, we propose the Combinatorial Neural Network Training Algorithm (CoNNTrA), that leverages a coordinate gradient descent-based approach for training deep learning models with finite discrete learning parameters. Next, we elaborate on the theoretical underpinnings and evaluate the computational complexity of CoNNTrA. As a proof of concept, we use CoNNTrA to train deep learning models with ternary learning parameters on the MNIST, Iris and ImageNet data sets and compare their performance to the same models trained using Backpropagation. We use following performance metrics for the comparison: (i) Training error; (ii) Validation error; (iii) Memory usage; and (iv) Training time. Our results indicate that CoNNTrA models use 32x less memory and have errors at par with the Backpropagation models.
Deep learning based methods have seen a massive rise in popularity for hyperspectral image classification over the past few years. However, the success of deep learning is attributed greatly to numerous labeled samples. It is still very challenging to use only a few labeled samples to train deep learning models to reach a high classification accuracy. An active deep-learning framework trained by an end-to-end manner is, therefore, proposed by this paper in order to minimize the hyperspectral image classification costs. First, a deep densely connected convolutional network is considered for hyperspectral image classification. Different from the traditional active learning methods, an additional network is added to the designed deep densely connected convolutional network to predict the loss of input samples. Then, the additional network could be used to suggest unlabeled samples that the deep densely connected convolutional network is more likely to produce a wrong label. Note that the additional network uses the intermediate features of the deep densely connected convolutional network as input. Therefore, the proposed method is an end-to-end framework. Subsequently, a few of the selected samples are labelled manually and added to the training samples. The deep densely connected convolutional network is therefore trained using the new training set. Finally, the steps above are repeated to train the whole framework iteratively. Extensive experiments illustrates that the method proposed could reach a high accuracy in classification after selecting just a few samples.
To conduct a radiomics or deep learning research experiment, the radiologists or physicians need to grasp the needed programming skills, which, however, could be frustrating and costly when they have limited coding experience. In this paper, we present DARWIN, a flexible research platform with a graphical user interface for medical imaging research. Our platform is consists of a radiomics module and a deep learning module. The radiomics module can extract more than 1000 dimension features(first-, second-, and higher-order) and provided many draggable supervised and unsupervised machine learning models. Our deep learning module integrates state of the art architectures of classification, detection, and segmentation tasks. It allows users to manually select hyperparameters, or choose an algorithm to automatically search for the best ones. DARWIN also offers the possibility for users to define a custom pipeline for their experiment. These flexibilities enable radiologists to carry out various experiments easily.
Extracting behavioral measurements non-invasively from video is stymied by the fact that it is a hard computational problem. Recent advances in deep learning have tremendously advanced predicting posture from videos directly, which quickly impacted neuroscience and biology more broadly. In this primer we review the budding field of motion capture with deep learning. In particular, we will discuss the principles of those novel algorithms, highlight their potential as well as pitfalls for experimentalists, and provide a glimpse into the future.
The performance of face analysis and recognition systems depends on the quality of the acquired face data, which is influenced by numerous factors. Automatically assessing the quality of face data in terms of biometric utility can thus be useful to detect low-quality data and make decisions accordingly. This survey provides an overview of the face image quality assessment literature, which predominantly focuses on visible wavelength face image input. A trend towards deep learning based methods is observed, including notable conceptual differences among the recent approaches, such as the integration of quality assessment into face recognition models. Besides image selection, face image quality assessment can also be used in a variety of other application scenarios, which are discussed herein. Open issues and challenges are pointed out, i.a. highlighting the importance of comparability for algorithm evaluations, and the challenge for future work to create deep learning approaches that are interpretable in addition to providing accurate utility predictions.
Deep learning has become the workhorse for a wide range of natural language processing applications. But much of the success of deep learning relies on annotated examples. Annotation is time-consuming and expensive to produce at scale. Here we are interested in methods for reducing the required quantity of annotated data -- by making the learning methods more knowledge efficient so as to make them more applicable in low annotation (low resource) settings. There are various classical approaches to making the models more knowledge efficient such as multi-task learning, transfer learning, weakly supervised and unsupervised learning etc. This thesis focuses on adapting such classical methods to modern deep learning models and algorithms. This thesis describes four works aimed at making machine learning models more knowledge efficient. First, we propose a knowledge rich deep learning model (KRDL) as a unifying learning framework for incorporating prior knowledge into deep models. In particular, we apply KRDL built on Markov logic networks to denoise weak supervision. Second, we apply a KRDL model to assist the machine reading models to find the correct evidence sentences that can support their decision. Third, we investigate the knowledge transfer techniques in multilingual setting, where we proposed a method that can improve pre-trained multilingual BERT based on the bilingual dictionary. Fourth, we present an episodic memory network for language modelling, in which we encode the large external knowledge for the pre-trained GPT.
Consumer electronic devices such as mobile handsets, goods tagged with RFID labels, location and position sensors are continuously generating a vast amount of location enriched data called geospatial data. Conventionally such geospatial data is used for military applications. In recent times, many useful civilian applications have been designed and deployed around such geospatial data. For example, a recommendation system to suggest restaurants or places of attraction to a tourist visiting a particular locality. At the same time, civic bodies are harnessing geospatial data generated through remote sensing devices to provide better services to citizens such as traffic monitoring, pothole identification, and weather reporting. Typically such applications are leveraged upon non-hierarchical machine learning techniques such as Naive-Bayes Classifiers, Support Vector Machines, and decision trees. Recent advances in the field of deep-learning showed that Neural Network-based techniques outperform conventional techniques and provide effective solutions for many geospatial data analysis tasks such as object recognition, image classification, and scene understanding. The chapter presents a survey on the current state of the applications of deep learning techniques for analyzing geospatial data. The chapter is organized as below: (i) A brief overview of deep learning algorithms. (ii)Geospatial Analysis: a Data Science Perspective (iii) Deep-learning techniques for Remote Sensing data analytics tasks (iv) Deep-learning techniques for GPS data analytics(iv) Deep-learning techniques for RFID data analytics.
Vehicle route prediction is one of the significant tasks in vehicles mobility. It is one of the means to reduce the accidents and increase comfort in human life. The task of route prediction becomes simpler with the development of certain machine learning and deep learning libraries. Meanwhile, the security and privacy issues are always lying in the vehicle communication as well as in route prediction. Therefore, we proposed a framework which will reduce these issues in vehicle communication and predict the route of vehicles in crossroads. Specifically, our proposed framework consists of two modules and both are working in sequence. The first module of our framework using a deep learning for recognizing the vehicle license plate number. Then, the second module using supervised learning algorithm of machine learning for predicting the route of the vehicle by using velocity difference and previous mobility patterns as the features of machine learning algorithm. Experiment results shows that accuracy of our framework.
Deep learning-based image reconstruction methods have achieved promising results across multiple MRI applications. However, most approaches require large-scale fully-sampled ground truth data for supervised training. Acquiring fully-sampled data is often either difficult or impossible, particularly for dynamic contrast enhancement (DCE), 3D cardiac cine, and 4D flow. We present a deep learning framework for MRI reconstruction without any fully-sampled data using generative adversarial networks. We test the proposed method in two scenarios: retrospectively undersampled fast spin echo knee exams and prospectively undersampled abdominal DCE. The method recovers more anatomical structure compared to conventional methods.
I propose a state of the art deep neural architectural solution for handwritten character recognition for Bengali alphabets, compound characters as well as numerical digits that achieves state-of-the-art accuracy 96.8% in just 11 epochs. Similar work has been done before by Chatterjee, Swagato, et al. but they achieved 96.12% accuracy in about 47 epochs. The deep neural architecture used in that paper was fairly large considering the inclusion of the weights of the ResNet 50 model which is a 50 layer Residual Network. This proposed model achieves higher accuracy as compared to any previous work & in a little number of epochs. ResNet50 is a good model trained on the ImageNet dataset, but I propose an HCR network that is trained from the scratch on Bengali characters without the "Ensemble Learning" that can outperform previous architectures.
Recently, deep learning approaches for accelerated MRI have been extensively studied thanks to their high performance reconstruction in spite of significantly reduced runtime complexity. These neural networks are usually trained in a supervised manner, so matched pairs of subsampled and fully sampled k-space data are required. Unfortunately, it is often difficult to acquire matched fully sampled k-space data, since the acquisition of fully sampled k-space data requires long scan time and often leads to the change of the acquisition protocol. Therefore, unpaired deep learning without matched label data has become a very important research topic. In this paper, we propose an unpaired deep learning approach using a optimal transport driven cycle-consistent generative adversarial network (OT-cycleGAN) that employs a single pair of generator and discriminator. The proposed OT-cycleGAN architecture is rigorously derived from a dual formulation of the optimal transport formulation using a specially designed penalized least squares cost. The experimental results show that our method can reconstruct high resolution MR images from accelerated k- space data from both single and multiple coil acquisition, without requiring matched reference data.
This paper develops a deep learning framework based on convolutional neural networks (CNNs) that enable real-time extraction of full-field subpixel structural displacements from videos. In particular, two new CNN architectures are designed and trained on a dataset generated by the phase-based motion extraction method from a single lab-recorded high-speed video of a dynamic structure. As displacement is only reliable in the regions with sufficient texture contrast, the sparsity of motion field induced by the texture mask is considered via the network architecture design and loss function definition. Results show that, with the supervision of full and sparse motion field, the trained network is capable of identifying the pixels with sufficient texture contrast as well as their subpixel motions. The performance of the trained networks is tested on various videos of other structures to extract the full-field motion (e.g., displacement time histories), which indicates that the trained networks have generalizability to accurately extract full-field subtle displacements for pixels with sufficient texture contrast.
We utilize classical facts from topology to show that the classification problem in machine learning is always solvable under very mild conditions. Furthermore, we show that a softmax classification network acts on an input topological space by a finite sequence of topological moves to achieve the classification task. Moreover, given a training dataset, we show how topological formalism can be used to suggest the appropriate architectural choices for neural networks designed to be trained as classifiers on the data. Finally, we show how the architecture of a neural network cannot be chosen independently from the shape of the underlying data. To demonstrate these results, we provide example datasets and show how they are acted upon by neural nets from this topological perspective.
Embryo quality assessment after in vitro fertilization (IVF) is primarily done visually by embryologists. Variability among assessors, however, remains one of the main causes of the low success rate of IVF. This study aims to develop an automated embryo assessment based on a deep learning model. This study includes a total of 1084 images from 1226 embryos. The images were captured by an inverted microscope at day 3 after fertilization. The images were labelled based on Veeck criteria that differentiate embryos to grade 1 to 5 based on the size of the blastomere and the grade of fragmentation. Our deep learning grading results were compared to the grading results from trained embryologists to evaluate the model performance. Our best model from fine-tuning a pre-trained ResNet50 on the dataset results in 91.79% accuracy. The model presented could be developed into an automated embryo assessment method in point-of-care settings.
This paper gives a detailed review of reinforcement learning (RL) in combinatorial optimization, introduces the history of combinatorial optimization starting in the 1950s, and compares it with the RL algorithms of recent years. This paper explicitly looks at a famous combinatorial problem-traveling salesperson problem (TSP). It compares the approach of modern RL algorithms for the TSP with an approach published in the 1970s. By comparing the similarities and variances between these methodologies, the paper demonstrates how RL algorithms are optimized due to the evolution of machine learning techniques and computing power. The paper then briefly introduces the deep learning approach to the TSP named deep RL, which is an extension of the traditional mathematical framework. In deep RL, attention and feature encoding mechanisms are introduced to generate near-optimal solutions. The survey shows that integrating the deep learning mechanism, such as attention with RL, can effectively approximate the TSP. The paper also argues that deep learning could be a generic approach that can be integrated with any traditional RL algorithm to enhance the outcomes of the TSP.