Cutting-edge diffusion models produce images with high quality and customizability, enabling them to be used for commercial art and graphic design purposes. But do diffusion models create unique works of art, or are they stealing content directly from their training sets? In this work, we study image retrieval frameworks that enable us to compare generated images with training samples and detect when content has been replicated. Applying our frameworks to diffusion models trained on multiple datasets including Oxford flowers, Celeb-A, ImageNet, and LAION, we discuss how factors such as training set size impact rates of content replication. We also identify cases where diffusion models, including the popular Stable Diffusion model, blatantly copy from their training data.
In reinforcement learning applications like robotics, agents usually need to deal with various input/output features when specified with different state/action spaces by their developers or physical restrictions. This indicates unnecessary re-training from scratch and considerable sample inefficiency, especially when agents follow similar solution steps to achieve tasks. In this paper, we aim to transfer similar high-level goal-transition knowledge to alleviate the challenge. Specifically, we propose PILoT, i.e., Planning Immediate Landmarks of Targets. PILoT utilizes the universal decoupled policy optimization to learn a goal-conditioned state planner; then, distills a goal-planner to plan immediate landmarks in a model-free style that can be shared among different agents. In our experiments, we show the power of PILoT on various transferring challenges, including few-shot transferring across action spaces and dynamics, from low-dimensional vector states to image inputs, from simple robot to complicated morphology; and we also illustrate a zero-shot transfer solution from a simple 2D navigation task to the harder Ant-Maze task.
Current self-supervised learning algorithms are often modality-specific and require large amounts of computational resources. To address these issues, we increase the training efficiency of data2vec, a learning objective that generalizes across several modalities. We do not encode masked tokens, use a fast convolutional decoder and amortize the effort to build teacher representations. data2vec 2.0 benefits from the rich contextualized target representations introduced in data2vec which enable a fast self-supervised learner. Experiments on ImageNet-1K image classification show that data2vec 2.0 matches the accuracy of Masked Autoencoders in 16.4x lower pre-training time, on Librispeech speech recognition it performs as well as wav2vec 2.0 in 10.6x less time, and on GLUE natural language understanding it matches a retrained RoBERTa model in half the time. Trading some speed for accuracy results in ImageNet-1K top-1 accuracy of 86.8\% with a ViT-L model trained for 150 epochs.
Class distribution plays an important role in learning deep classifiers. When the proportion of each class in the test set differs from the training set, the performance of classification nets usually degrades. Such a label distribution shift problem is common in medical diagnosis since the prevalence of disease vary over location and time. In this paper, we propose the first method to tackle label shift for medical image classification, which effectively adapt the model learned from a single training label distribution to arbitrary unknown test label distribution. Our approach innovates distribution calibration to learn multiple representative classifiers, which are capable of handling different one-dominating-class distributions. When given a test image, the diverse classifiers are dynamically aggregated via the consistency-driven test-time adaptation, to deal with the unknown test label distribution. We validate our method on two important medical image classification tasks including liver fibrosis staging and COVID-19 severity prediction. Our experiments clearly show the decreased model performance under label shift. With our method, model performance significantly improves on all the test datasets with different label shifts for both medical image diagnosis tasks.
The capability to detect boulders on the surface of small bodies is beneficial for vision-based applications such as navigation and hazard detection during critical operations. This task is challenging due to the wide assortment of irregular shapes, the characteristics of the boulders population, and the rapid variability in the illumination conditions. The authors address this challenge by designing a multi-step training approach to develop a data-driven image processing pipeline to robustly detect and segment boulders scattered over the surface of a small body. Due to the limited availability of labeled image-mask pairs, the developed methodology is supported by two artificial environments designed in Blender specifically for this work. These are used to generate a large amount of synthetic image-label sets, which are made publicly available to the image processing community. The methodology presented addresses the challenges of varying illumination conditions, irregular shapes, fast training time, extensive exploration of the architecture design space, and domain gap between synthetic and real images from previously flown missions. The performance of the developed image processing pipeline is tested both on synthetic and real images, exhibiting good performances, and high generalization capabilities
In ophthalmological imaging, multiple imaging systems, such as color fundus, infrared, fluorescein angiography, optical coherence tomography (OCT) or OCT angiography, are often involved to make a diagnosis of retinal disease. Multi-modal retinal registration techniques can assist ophthalmologists by providing a pixel-based comparison of aligned vessel structures in images from different modalities or acquisition times. To this end, we propose an end-to-end trainable deep learning method for multi-modal retinal image registration. Our method extracts convolutional features from the vessel structure for keypoint detection and description and uses a graph neural network for feature matching. The keypoint detection and description network and graph neural network are jointly trained in a self-supervised manner using synthetic multi-modal image pairs and are guided by synthetically sampled ground truth homographies. Our method demonstrates higher registration accuracy as competing methods for our synthetic retinal dataset and generalizes well for our real macula dataset and a public fundus dataset.
Event cameras such as DAVIS can simultaneously output high temporal resolution events and low frame-rate intensity images, which own great potential in capturing scene motion, such as optical flow estimation. Most of the existing optical flow estimation methods are based on two consecutive image frames and can only estimate discrete flow at a fixed time interval. Previous work has shown that continuous flow estimation can be achieved by changing the quantities or time intervals of events. However, they are difficult to estimate reliable dense flow , especially in the regions without any triggered events. In this paper, we propose a novel deep learning-based dense and continuous optical flow estimation framework from a single image with event streams, which facilitates the accurate perception of high-speed motion. Specifically, we first propose an event-image fusion and correlation module to effectively exploit the internal motion from two different modalities of data. Then we propose an iterative update network structure with bidirectional training for optical flow prediction. Therefore, our model can estimate reliable dense flow as two-frame-based methods, as well as estimate temporal continuous flow as event-based methods. Extensive experimental results on both synthetic and real captured datasets demonstrate that our model outperforms existing event-based state-of-the-art methods and our designed baselines for accurate dense and continuous optical flow estimation.
Multi-instance learning (MIL) is a great paradigm for dealing with complex data and has achieved impressive achievements in a number of fields, including image classification, video anomaly detection, and far more. Each data sample is referred to as a bag containing several unlabeled instances, and the supervised information is only provided at the bag-level. The safety of MIL learners is concerning, though, as we can greatly fool them by introducing a few adversarial perturbations. This can be fatal in some cases, such as when users are unable to access desired images and criminals are attempting to trick surveillance cameras. In this paper, we design two adversarial perturbations to interpret the vulnerability of MIL methods. The first method can efficiently generate the bag-specific perturbation (called customized) with the aim of outsiding it from its original classification region. The second method builds on the first one by investigating the image-agnostic perturbation (called universal) that aims to affect all bags in a given data set and obtains some generalizability. We conduct various experiments to verify the performance of these two perturbations, and the results show that both of them can effectively fool MIL learners. We additionally propose a simple strategy to lessen the effects of adversarial perturbations. Source codes are available at https://github.com/InkiInki/MI-UAP.
The cover is the face of a book and is a point of attraction for the readers. Designing book covers is an essential task in the publishing industry. One of the main challenges in creating a book cover is representing the theme of the book's content in a single image. In this research, we explore ways to produce a book cover using artificial intelligence based on the fact that there exists a relationship between the summary of the book and its cover. Our key motivation is the application of text-to-image synthesis methods to generate images from given text or captions. We explore several existing text-to-image conversion techniques for this purpose and propose an approach to exploit these frameworks for producing book covers from provided summaries. We construct a dataset of English books that contains a large number of samples of summaries of existing books and their cover images. In this paper, we describe our approach to collecting, organizing, and pre-processing the dataset to use it for training models. We apply different text-to-image synthesis techniques to generate book covers from the summary and exhibit the results in this paper.
Although augmentations (e.g., perturbation of graph edges, image crops) boost the efficiency of Contrastive Learning (CL), feature level augmentation is another plausible, complementary yet not well researched strategy. Thus, we present a novel spectral feature argumentation for contrastive learning on graphs (and images). To this end, for each data view, we estimate a low-rank approximation per feature map and subtract that approximation from the map to obtain its complement. This is achieved by the proposed herein incomplete power iteration, a non-standard power iteration regime which enjoys two valuable byproducts (under mere one or two iterations): (i) it partially balances spectrum of the feature map, and (ii) it injects the noise into rebalanced singular values of the feature map (spectral augmentation). For two views, we align these rebalanced feature maps as such an improved alignment step can focus more on less dominant singular values of matrices of both views, whereas the spectral augmentation does not affect the spectral angle alignment (singular vectors are not perturbed). We derive the analytical form for: (i) the incomplete power iteration to capture its spectrum-balancing effect, and (ii) the variance of singular values augmented implicitly by the noise. We also show that the spectral augmentation improves the generalization bound. Experiments on graph/image datasets show that our spectral feature augmentation outperforms baselines, and is complementary with other augmentation strategies and compatible with various contrastive losses.