Convolutional neural networks (CNNs) are a promising technique for automated glaucoma diagnosis from images of the fundus, and these images are routinely acquired as part of an ophthalmic exam. Nevertheless, CNNs typically require a large amount of well-labeled data for training, which may not be available in many biomedical image classification applications, especially when diseases are rare and where labeling by experts is costly. This article makes two contributions to address this issue: 1) It extends the conventional Siamese network and introduces a training method for low-shot learning when labeled data are limited and imbalanced, and 2) it introduces a novel semi-supervised learning strategy that uses additional unlabeled training data to achieve greater accuracy. Our proposed multi-task Siamese network (MTSN) can employ any backbone CNN, and we demonstrate with four backbone CNNs that its accuracy with limited training data approaches the accuracy of backbone CNNs trained with a dataset that is 50 times larger. We also introduce One-Vote Veto (OVV) self-training, a semi-supervised learning strategy that is designed specifically for MTSNs. By taking both self-predictions and contrastive predictions of the unlabeled training data into account, OVV self-training provides additional pseudo labels for fine-tuning a pre-trained MTSN. Using a large (imbalanced) dataset with 66,715 fundus photographs acquired over 15 years, extensive experimental results demonstrate the effectiveness of low-shot learning with MTSN and semi-supervised learning with OVV self-training. Three additional, smaller clinical datasets of fundus images acquired under different conditions (cameras, instruments, locations, populations) are used to demonstrate the generalizability of the proposed methods.

To investigate the diagnostic accuracy of deep learning (DL) algorithms to detect primary open-angle glaucoma (POAG) trained on fundus photographs from the Ocular Hypertension Treatment Study (OHTS). 66,715 photographs from 3,272 eyes were used to train and test a ResNet-50 model to detect the OHTS Endpoint Committee POAG determination based on optic disc (n=287 eyes, 3,502 photographs) and/or visual field (n=198 eyes, 2,300 visual fields) changes. OHTS training, validation and testing sets were randomly determined using an 85-5-10 percentage split by subject. Three independent test sets were used to estimate the generalizability of the model: UCSD Diagnostic Innovations in Glaucoma Study (DIGS, USA), ACRIMA (Spain) and Large-scale Attention-based Glaucoma (LAG, China). The DL model achieved an AUROC (95% CI) of 0.88 (0.82, 0.92) for the overall OHTS POAG endpoint. For the OHTS endpoints based on optic disc changes or visual field changes, AUROCs were 0.91 (0.88, 0.94) and 0.86 (0.76, 0.93), respectively. False-positive rates (at 90% specificity) were higher in photographs of eyes that later developed POAG by disc or visual field (19.1%) compared to eyes that did not develop POAG (7.3%) during their OHTS follow-up. The diagnostic accuracy of the DL model developed on the OHTS optic disc endpoint applied to 3 independent datasets was lower with AUROC ranging from 0.74 to 0.79. High diagnostic accuracy of the current model suggests that DL can be used to automate the determination of POAG for clinical trials and management. In addition, the higher false-positive rate in early photographs of eyes that later developed POAG suggests that DL models detected POAG in some eyes earlier than the OHTS Endpoint Committee.

Purpose: To compare the diagnostic accuracy and explainability of a new Vision Transformer deep learning technique, Data-efficient image Transformer (DeiT), and Resnet-50, trained on fundus photographs from the Ocular Hypertension Treatment Study (OHTS) to detect primary open-angle glaucoma (POAG) and to identify the salient areas of the photographs most important for each model's decision-making process. Study Design: Evaluation of a diagnostic technology Subjects, Participants, and/or Controls: 66,715 photographs from 1,636 OHTS participants and an additional five external datasets of 16137 photographs of healthy and glaucoma eyes. Methods, Intervention, or Testing: DeiT models were trained to detect five ground truth OHTS POAG classifications: OHTS Endpoint Committee POAG determinations due to disc changes (Model 1), visual field changes (Model 2), or either disc or visual field changes (Model 3) and reading center determinations based on disc (Model 4) and visual fields (Model 5). The best-performing DeiT models were compared to ResNet-50 on OHTS and five external datasets. Main Outcome Measures: Diagnostic performance was compared using areas under the receiver operating characteristic curve (AUROC) and sensitivities at fixed specificities. The explainability of the DeiT and ResNet-50 models was compared by evaluating the attention maps derived directly from DeiT to 3 gradient-weighted class activation map generation strategies. Results: Compared to our best-performing ResNet-50 models, the DeiT models demonstrated similar performance on the OHTS test sets for all five-ground truth POAG labels; AUROC ranged from 0.82 (Model 5) to 0.91 (Model 1). However, the AUROC of DeiT was consistently higher than ResNet-50 on the five external datasets. For example, AUROC for the main OHTS endpoint (Model 3) was between 0.08 and 0.20 higher in the DeiT compared to ResNet-50 models. The saliency maps from the DeiT highlight localized areas of the neuroretinal rim, suggesting the use of important clinical features for classification, while the same maps in the ResNet-50 models show a more diffuse, generalized distribution around the optic disc, Conclusions: Vision transformer has the potential to improve the generalizability and explainability of deep learning models for the detection of eye disease and possibly other medical conditions that rely on imaging modalities for clinical diagnosis and management.

CoralNet is a cloud-based website and platform for manual, semi-automatic and automatic analysis of coral reef images. Users access CoralNet through optimized web-based workflows for common tasks, and other systems can interface through API’s. Today, marine scientists are widely using CoralNet, and nearly 3,000 registered users have up-loaded 1,741,855 images from 2,040 distinct sources with over 65 million annotations. CoralNet is hosted on AWS, is free for users, and the code is open source 1. In January 2021, we released CoralNet 1.0 which has a new machine learning engine. This paper provides an overview of that engine, and the process of choosing the particular architecture, its training, and a comparison to some of the most promising architectures. In a nutshell, CoralNet 1.0 uses transfer learning with an EfficientNet-B0 backbone that is trained on 16M labelled patches from benthic images and a hierarchical Multi-layer Perceptron classifier that is trained on source-specific labelled data. When evaluated on a hold-out test set of 26 sources, the error rate of CoralNet 1.0 was 18.4% (relative) lower than CoralNet Beta.

We present Neural Reflectance Fields, a novel deep scene representation that encodes volume density, normal and reflectance properties at any 3D point in a scene using a fully-connected neural network. We combine this representation with a physically-based differentiable ray marching framework that can render images from a neural reflectance field under any viewpoint and light. We demonstrate that neural reflectance fields can be estimated from images captured with a simple collocated camera-light setup, and accurately model the appearance of real-world scenes with complex geometry and reflectance. Once estimated, they can be used to render photo-realistic images under novel viewpoint and (non-collocated) lighting conditions and accurately reproduce challenging effects like specularities, shadows and occlusions. This allows us to perform high-quality view synthesis and relighting that is significantly better than previous methods. We also demonstrate that we can compose the estimated neural reflectance field of a real scene with traditional scene models and render them using standard Monte Carlo rendering engines. Our work thus enables a complete pipeline from high-quality and practical appearance acquisition to 3D scene composition and rendering.

In this work, we address the problem of precisely localizing key frames of an action, for example, the precise time that a pitcher releases a baseball, or the precise time that a crowd begins to applaud. Key frame localization is a largely overlooked and important action-recognition problem, for example in the field of neuroscience, in which we would like to understand the neural activity that produces the start of a bout of an action. To address this problem, we introduce a novel structured loss function that properly weights the types of errors that matter in such applications: it more heavily penalizes extra and missed action start detections over small misalignments. Our structured loss is based on the best matching between predicted and labeled action starts. We train recurrent neural networks (RNNs) to minimize differentiable approximations of this loss. To evaluate these methods, we introduce the Mouse Reach Dataset, a large, annotated video dataset of mice performing a sequence of actions. The dataset was collected and labeled by experts for the purpose of neuroscience research. On this dataset, we demonstrate that our method outperforms related approaches and baseline methods using an unstructured loss.

We introduce a novel learning-based method to reconstruct the high-quality geometry and complex, spatially-varying BRDF of an arbitrary object from a sparse set of only six images captured by wide-baseline cameras under collocated point lighting. We first estimate per-view depth maps using a deep multi-view stereo network; these depth maps are used to coarsely align the different views. We propose a novel multi-view reflectance estimation network architecture that is trained to pool features from these coarsely aligned images and predict per-view spatially-varying diffuse albedo, surface normals, specular roughness and specular albedo. We do this by jointly optimizing the latent space of our multi-view reflectance network to minimize the photometric error between images rendered with our predictions and the input images. While previous state-of-the-art methods fail on such sparse acquisition setups, we demonstrate, via extensive experiments on synthetic and real data, that our method produces high-quality reconstructions that can be used to render photorealistic images.

We present a deep learning approach to reconstruct scene appearance from unstructured images captured under collocated point lighting. At the heart of Deep Reflectance Volumes is a novel volumetric scene representation consisting of opacity, surface normal and reflectance voxel grids. We present a novel physically-based differentiable volume ray marching framework to render these scene volumes under arbitrary viewpoint and lighting. This allows us to optimize the scene volumes to minimize the error between their rendered images and the captured images. Our method is able to reconstruct real scenes with challenging non-Lambertian reflectance and complex geometry with occlusions and shadowing. Moreover, it accurately generalizes to novel viewpoints and lighting, including non-collocated lighting, rendering photorealistic images that are significantly better than state-of-the-art mesh-based methods. We also show that our learned reflectance volumes are editable, allowing for modifying the materials of the captured scenes.

The task of Visual Commonsense Reasoning is extremely challenging in the sense that the model has to not only be able to answer a question given an image, but also be able to learn to reason. The baselines introduced in this task are quite limiting because two networks are trained for predicting answers and rationales separately. Question and image is used as input to train answer prediction network while question, image and correct answer are used as input in the rationale prediction network. As rationale is conditioned on the correct answer, it is based on the assumption that we can solve Visual Question Answering task without any error - which is over ambitious. Moreover, such an approach makes both answer and rationale prediction two completely independent VQA tasks rendering cognition task meaningless. In this paper, we seek to address these issues by proposing an end-to-end trainable model which considers both answers and their reasons jointly. Specifically, we first predict the answer for the question and then use the chosen answer to predict the rationale. However, a trivial design of such a model becomes non-differentiable which makes it difficult to train. We solve this issue by proposing four approaches - softmax, gumbel-softmax, reinforcement learning based sampling and direct cross entropy against all pairs of answers and rationales. We demonstrate through experiments that our model performs competitively against current state-of-the-art. We conclude with an analysis of presented approaches and discuss avenues for further work.

Imaging through dynamic refractive media, such as looking into turbulent water, or through hot air, is challenging since light rays are bent by unknown amounts leading to complex geometric distortions. Inverting these distortions and recovering high quality images is an inherently ill-posed problem, leading previous works to require extra information such as high frame-rate video or a template image, which limits their applicability in practice. This paper proposes training a deep convolution neural network to undistort dynamic refractive effects using only a single image. The neural network is able to solve this ill-posed problem by learning image priors as well as distortion priors. Our network consists of two parts, a warping net to remove geometric distortion and a color predictor net to further refine the restoration. Adversarial loss is used to achieve better visual quality and help the network hallucinate missing and blurred information. To train our network, we collect a large training set of images distorted by a turbulent water surface. Unlike prior works on water undistortion, our method is trained end-to-end, only requires a single image and does not use a ground truth template at test time. Experiments show that by exploiting the structure of the problem, our network outperforms state-of-the-art deep image to image translation.

We propose a general framework for unsupervised domain adaptation, which allows deep neural networks trained on a source domain to be tested on a different target domain without requiring any training annotations in the target domain. This is achieved by adding extra networks and losses that help regularize the features extracted by the backbone encoder network. To this end we propose the novel use of the recently proposed unpaired image-toimage translation framework to constrain the features extracted by the encoder network. Specifically, we require that the features extracted are able to reconstruct the images in both domains. In addition we require that the distribution of features extracted from images in the two domains are indistinguishable. Many recent works can be seen as specific cases of our general framework. We apply our method for domain adaptation between MNIST, USPS, and SVHN datasets, and Amazon, Webcam and DSLR Office datasets in classification tasks, and also between GTA5 and Cityscapes datasets for a segmentation task. We demonstrate state of the art performance on each of these datasets.

Traditional imaging methods and computer vision algorithms are often ineffective when images are acquired in scattering media, such as underwater, fog, and biological tissue. Here, we explore the use of light field imaging and algorithms for image restoration and depth estimation that address the image degradation from the medium. Towards this end, we make the following three contributions. First, we present a new single image restoration algorithm which removes backscatter and attenuation from images better than existing methods do, and apply it to each view in the light field. Second, we combine a novel transmission based depth cue with existing correspondence and defocus cues to improve light field depth estimation. In densely scattering media, our transmission depth cue is critical for depth estimation since the images have low signal to noise ratios which significantly degrades the performance of the correspondence and defocus cues. Finally, we propose shearing and refocusing multiple views of the light field to recover a single image of higher quality than what is possible from a single view. We demonstrate the benefits of our method through extensive experimental results in a water tank.

In this work, we present a novel 3D-Convolutional Neural Network (CNN) architecture called I2I-3D that predicts boundary location in volumetric data. Our fine-to-fine, deeply supervised framework addresses three critical issues to 3D boundary detection: (1) efficient, holistic, end-to-end volumetric label training and prediction (2) precise voxel-level prediction to capture fine scale structures prevalent in medical data and (3) directed multi-scale, multi-level feature learning. We evaluate our approach on a dataset consisting of 93 medical image volumes with a wide variety of anatomical regions and vascular structures. In the process, we also introduce HED-3D, a 3D extension of the state-of-the-art 2D edge detector (HED). We show that our deep learning approach out-performs, the current state-of-the-art in 3D vascular boundary detection (structured forests 3D), by a large margin, as well as HED applied to slices, and HED-3D while successfully localizing fine structures. With our approach, boundary detection takes about one minute on a typical 512x512x512 volume.

Photometric stereo is widely used for 3D reconstruction. However, its use in scattering media such as water, biological tissue and fog has been limited until now, because of forward scattered light from both the source and object, as well as light scattered back from the medium (backscatter). Here we make three contributions to address the key modes of light propagation, under the common single scattering assumption for dilute media. First, we show through extensive simulations that single-scattered light from a source can be approximated by a point light source with a single direction. This alleviates the need to handle light source blur explicitly. Next, we model the blur due to scattering of light from the object. We measure the object point-spread function and introduce a simple deconvolution method. Finally, we show how imaging fluorescence emission where available, eliminates the backscatter component and increases the signal-to-noise ratio. Experimental results in a water tank, with different concentrations of scattering media added, show that deconvolution produces higher-quality 3D reconstructions than previous techniques, and that when combined with fluorescence, can produce results similar to that in clear water even for highly turbid media.

Global climate change and other anthropogenic stressors have heightened the need to rapidly characterize ecological changes in marine benthic communities across large scales. Digital photography enables rapid collection of survey images to meet this need, but the subsequent image annotation is typically a time consuming, manual task. We investigated the feasibility of using automated point-annotation to expedite cover estimation of the 17 dominant benthic categories from survey-images captured at four Pacific coral reefs. Inter- and intra- annotator variability among six human experts was quantified and compared to semi- and fully- automated annotation methods, which are made available at coralnet.ucsd.edu. Our results indicate high expert agreement for identification of coral genera, but lower agreement for algal functional groups, in particular between turf algae and crustose coralline algae. This indicates the need for unequivocal definitions of algal groups, careful training of multiple annotators, and enhanced imaging technology. Semi-automated annotation, where 50% of the annotation decisions were performed automatically, yielded cover estimate errors comparable to those of the human experts. Furthermore, fully-automated annotation yielded rapid, unbiased cover estimates but with increased variance. These results show that automated annotation can increase spatial coverage and decrease time and financial outlay for image-based reef surveys.