Magnetic Resonance Imaging (MRI) is a powerful medical imaging modality, but unfortunately suffers from long scan times which, aside from increasing operational costs, can lead to image artifacts due to patient motion. Motion during the acquisition leads to inconsistencies in measured data that manifest as blurring and ghosting if unaccounted for in the image reconstruction process. Various deep learning based reconstruction techniques have been proposed which decrease scan time by reducing the number of measurements needed for a high fidelity reconstructed image. Additionally, deep learning has been used to correct motion using end-to-end techniques. This, however, increases susceptibility to distribution shifts at test time (sampling pattern, motion level). In this work we propose a framework for jointly reconstructing highly sub-sampled MRI data while estimating patient motion using score-based generative models. Our method does not make specific assumptions on the sampling trajectory or motion pattern at training time and thus can be flexibly applied to various types of measurement models and patient motion. We demonstrate our framework on retrospectively accelerated 2D brain MRI corrupted by rigid motion.
This paper proposes a new technique based on a non-linear Minmax Detector Based (MDB) filter for image restoration. The aim of image enhancement is to reconstruct the true image from the corrupted image. The process of image acquisition frequently leads to degradation and the quality of the digitized image becomes inferior to the original image. Image degradation can be due to the addition of different types of noise in the original image. Image noise can be modelled of many types and impulse noise is one of them. Impulse noise generates pixels with gray value not consistent with their local neighbourhood. It appears as a sprinkle of both light and dark or only light spots in the image. Filtering is a technique for enhancing the image. Linear filter is the filtering in which the value of an output pixel is a linear combination of neighborhood values, which can produce blur in the image. Thus a variety of smoothing techniques have been developed that are non linear. Median filter is the one of the most popular non-linear filter. When considering a small neighborhood it is highly efficient but for large window and in case of high noise it gives rise to more blurring to image. The Centre Weighted Mean (CWM) filter has got a better average performance over the median filter. However the original pixel corrupted and noise reduction is substantial under high noise condition. Hence this technique has also blurring affect on the image. To illustrate the superiority of the proposed approach, the proposed new scheme has been simulated along with the standard ones and various restored performance measures have been compared.
Single-pixel imaging (SPI) is a novel imaging technique whose working principle is based on the compressive sensing (CS) theory. In SPI, data is obtained through a series of compressive measurements and the corresponding image is reconstructed. Typically, the reconstruction algorithm such as basis pursuit relies on the sparsity assumption in images. However, recent advances in deep learning have found its uses in reconstructing CS images. Despite showing a promising result in simulations, it is often unclear how such an algorithm can be implemented in an actual SPI setup. In this paper, we demonstrate the use of deep learning on the reconstruction of SPI images in conjunction with block compressive sensing (BCS). We also proposed a novel reconstruction model based on convolutional neural networks that outperforms other competitive CS reconstruction algorithms. Besides, by incorporating BCS in our deep learning model, we were able to reconstruct images of any size above a certain smallest image size. In addition, we show that our model is capable of reconstructing images obtained from an SPI setup while being priorly trained on natural images, which can be vastly different from the SPI images. This opens up opportunity for the feasibility of pretrained deep learning models for CS reconstructions of images from various domain areas.
Homography estimation is erroneous in the case of large-baseline due to the low image overlay and limited receptive field. To address it, we propose a progressive estimation strategy by converting large-baseline homography into multiple intermediate ones, cumulatively multiplying these intermediate items can reconstruct the initial homography. Meanwhile, a semi-supervised homography identity loss, which consists of two components: a supervised objective and an unsupervised objective, is introduced. The first supervised loss is acting to optimize intermediate homographies, while the second unsupervised one helps to estimate a large-baseline homography without photometric losses. To validate our method, we propose a large-scale dataset that covers regular and challenging scenes. Experiments show that our method achieves state-of-the-art performance in large-baseline scenes while keeping competitive performance in small-baseline scenes. Code and dataset are available at https://github.com/megvii-research/LBHomo.
Image restoration is a challenging and ill-posed problem which also has been a long-standing issue. However, most of learning based restoration methods are proposed to target one degradation type which means they are lack of generalization. In this paper, we proposed a multi-branch restoration model inspired from the Human Visual System (i.e., Retinal Ganglion Cells) which can achieve multiple restoration tasks in a general framework. The experiments show that the proposed multi-branch architecture, called CMFNet, has competitive performance results on four datasets, including image dehazing, deraindrop, and deblurring, which are very common applications for autonomous cars. The source code and pretrained models of three restoration tasks are available at https://github.com/FanChiMao/CMFNet.
Puzzle solving is a combinatorial challenge due to the difficulty of matching adjacent pieces. Instead, we infer a mental image from all pieces, which a given piece can then be matched against avoiding the combinatorial explosion. Exploiting advancements in Generative Adversarial methods, we learn how to reconstruct the image given a set of unordered pieces, allowing the model to learn a joint embedding space to match an encoding of each piece to the cropped layer of the generator. Therefore we frame the problem as a R@1 retrieval task, and then solve the linear assignment using differentiable Hungarian attention, making the process end-to-end. In doing so our model is puzzle size agnostic, in contrast to prior deep learning methods which are single size. We evaluate on two new large-scale datasets, where our model is on par with deep learning methods, while generalizing to multiple puzzle sizes.
In the field of medical image, deep convolutional neural networks(ConvNets) have achieved great success in the classification, segmentation, and registration tasks thanks to their unparalleled capacity to learn image features. However, these tasks often require large amounts of manually annotated data and are labor-intensive. Therefore, it is of significant importance for us to study unsupervised semantic feature learning tasks. In our work, we propose to learn features in medical images by training ConvNets to recognize the geometric transformation applied to images and present a simple self-supervised task that can easily predict the geometric transformation. We precisely define a set of geometric transformations in mathematical terms and generalize this model to 3D, taking into account the distinction between spatial and time dimensions. We evaluated our self-supervised method on CMR images of different modalities (bSSFP, T2, LGE) and achieved accuracies of 96.4%, 97.5%, and 96.4%, respectively. The code and models of our paper will be published on: https://github.com/gaoxin492/Geometric_Transformation_CMR
Optoacoustic (OA) imaging is based on excitation of biological tissues with nanosecond-duration laser pulses followed by subsequent detection of ultrasound waves generated via light-absorption-mediated thermoelastic expansion. OA imaging features a powerful combination between rich optical contrast and high resolution in deep tissues. This enabled the exploration of a number of attractive new applications both in clinical and laboratory settings. However, no standardized datasets generated with different types of experimental set-up and associated processing methods are available to facilitate advances in broader applications of OA in clinical settings. This complicates an objective comparison between new and established data processing methods, often leading to qualitative results and arbitrary interpretations of the data. In this paper, we provide both experimental and synthetic OA raw signals and reconstructed image domain datasets rendered with different experimental parameters and tomographic acquisition geometries. We further provide trained neural networks to tackle three important challenges related to OA image processing, namely accurate reconstruction under limited view tomographic conditions, removal of spatial undersampling artifacts and anatomical segmentation for improved image reconstruction. Specifically, we define 18 experiments corresponding to the aforementioned challenges as benchmarks to be used as a reference for the development of more advanced processing methods.
In this work, we tackle the challenging problem of arbitrary image style transfer using a novel style feature representation learning method. A suitable style representation, as a key component in image stylization tasks, is essential to achieve satisfactory results. Existing deep neural network based approaches achieve reasonable results with the guidance from second-order statistics such as Gram matrix of content features. However, they do not leverage sufficient style information, which results in artifacts such as local distortions and style inconsistency. To address these issues, we propose to learn style representation directly from image features instead of their second-order statistics, by analyzing the similarities and differences between multiple styles and considering the style distribution. Specifically, we present Contrastive Arbitrary Style Transfer (CAST), which is a new style representation learning and style transfer method via contrastive learning. Our framework consists of three key components, i.e., a multi-layer style projector for style code encoding, a domain enhancement module for effective learning of style distribution, and a generative network for image style transfer. We conduct qualitative and quantitative evaluations comprehensively to demonstrate that our approach achieves significantly better results compared to those obtained via state-of-the-art methods. Code and models are available at https://github.com/zyxElsa/CAST_pytorch
Jitendra Malik once said, "Supervision is the opium of the AI researcher". Most deep learning techniques heavily rely on extreme amounts of human labels to work effectively. In today's world, the rate of data creation greatly surpasses the rate of data annotation. Full reliance on human annotations is just a temporary means to solve current closed problems in AI. In reality, only a tiny fraction of data is annotated. Annotation Efficient Learning (AEL) is a study of algorithms to train models effectively with fewer annotations. To thrive in AEL environments, we need deep learning techniques that rely less on manual annotations (e.g., image, bounding-box, and per-pixel labels), but learn useful information from unlabeled data. In this thesis, we explore five different techniques for handling AEL.