We present NeRFEditor, an efficient learning framework for 3D scene editing, which takes a video captured over 360{\deg} as input and outputs a high-quality, identity-preserving stylized 3D scene. Our method supports diverse types of editing such as guided by reference images, text prompts, and user interactions. We achieve this by encouraging a pre-trained StyleGAN model and a NeRF model to learn from each other mutually. Specifically, we use a NeRF model to generate numerous image-angle pairs to train an adjustor, which can adjust the StyleGAN latent code to generate high-fidelity stylized images for any given angle. To extrapolate editing to GAN out-of-domain views, we devise another module that is trained in a self-supervised learning manner. This module maps novel-view images to the hidden space of StyleGAN that allows StyleGAN to generate stylized images on novel views. These two modules together produce guided images in 360{\deg}views to finetune a NeRF to make stylization effects, where a stable fine-tuning strategy is proposed to achieve this. Experiments show that NeRFEditor outperforms prior work on benchmark and real-world scenes with better editability, fidelity, and identity preservation.
Prompt learning is one of the most effective and trending ways to adapt powerful vision-language foundation models like CLIP to downstream datasets by tuning learnable prompt vectors with very few samples. However, although prompt learning achieves excellent performance over in-domain data, it still faces the major challenge of generalizing to unseen classes and domains. Some existing prompt learning methods tackle this issue by adaptively generating different prompts for different tokens or domains but neglecting the ability of learned prompts to generalize to unseen domains. In this paper, we propose a novel prompt learning paradigm that directly generates domain invariant prompt generalizable to unseen domains, called MetaPrompt. Specifically, a dual-modality prompt tuning network is proposed to generate prompts for inputs from both image and text modalities. More importantly, we propose a meta-learning-based prompt tuning algorithm that explicitly constrains the prompt tuned on a specific domain or class also to achieve good performance on another domain or class. Extensive experiments on 11 datasets for base-to-new generalization and four datasets for domain generalization demonstrate that our method consistently and significantly outperforms existing methods.
Over the past few years, image-to-image (I2I) translation methods have been proposed to translate a given image into diverse outputs. Despite the impressive results, they mainly focus on the I2I translation between two domains, so the multi-domain I2I translation still remains a challenge. To address this problem, we propose a novel multi-domain unsupervised image-to-image translation (MDUIT) framework that leverages the decomposed content feature and appearance adaptive convolution to translate an image into a target appearance while preserving the given geometric content. We also exploit a contrast learning objective, which improves the disentanglement ability and effectively utilizes multi-domain image data in the training process by pairing the semantically similar images. This allows our method to learn the diverse mappings between multiple visual domains with only a single framework. We show that the proposed method produces visually diverse and plausible results in multiple domains compared to the state-of-the-art methods.
Attention-based neural networks, such as Transformers, have become ubiquitous in numerous applications, including computer vision, natural language processing, and time-series analysis. In all kinds of attention networks, the attention maps are crucial as they encode semantic dependencies between input tokens. However, most existing attention networks perform modeling or reasoning based on representations, wherein the attention maps of different layers are learned separately without explicit interactions. In this paper, we propose a novel and generic evolving attention mechanism, which directly models the evolution of inter-token relationships through a chain of residual convolutional modules. The major motivations are twofold. On the one hand, the attention maps in different layers share transferable knowledge, thus adding a residual connection can facilitate the information flow of inter-token relationships across layers. On the other hand, there is naturally an evolutionary trend among attention maps at different abstraction levels, so it is beneficial to exploit a dedicated convolution-based module to capture this process. Equipped with the proposed mechanism, the convolution-enhanced evolving attention networks achieve superior performance in various applications, including time-series representation, natural language understanding, machine translation, and image classification. Especially on time-series representation tasks, Evolving Attention-enhanced Dilated Convolutional (EA-DC-) Transformer outperforms state-of-the-art models significantly, achieving an average of 17% improvement compared to the best SOTA. To the best of our knowledge, this is the first work that explicitly models the layer-wise evolution of attention maps. Our implementation is available at https://github.com/pkuyym/EvolvingAttention
As an emerging approach to space situational awareness and space imaging, the practical use of an event-based camera in space imaging for precise source analysis is still in its infancy. The nature of event-based space imaging and data collection needs to be further explored to develop more effective event-based space image systems and advance the capabilities of event-based tracking systems with improved target measurement models. Moreover, for event measurements to be meaningful, a framework must be investigated for event-based camera calibration to project events from pixel array coordinates in the image plane to coordinates in a target resident space object's reference frame. In this paper, the traditional techniques of conventional astronomy are reconsidered to properly utilise the event-based camera for space imaging and space situational awareness. This paper presents the techniques and systems used for calibrating an event-based camera for reliable and accurate measurement acquisition. These techniques are vital in building event-based space imaging systems capable of real-world space situational awareness tasks. By calibrating sources detected using the event-based camera, the spatio-temporal characteristics of detected sources or `event sources' can be related to the photometric characteristics of the underlying astrophysical objects. Finally, these characteristics are analysed to establish a foundation for principled processing and observing techniques which appropriately exploit the capabilities of the event-based camera.
Masked image modeling (MIM) learns visual representation by masking and reconstructing image patches. Applying the reconstruction supervision on the CLIP representation has been proven effective for MIM. However, it is still under-explored how CLIP supervision in MIM influences performance. To investigate strategies for refining the CLIP-targeted MIM, we study two critical elements in MIM, i.e., the supervision position and the mask ratio, and reveal two interesting perspectives, relying on our developed simple pipeline, context autodecoder with CLIP target (CAE v2). Firstly, we observe that the supervision on visible patches achieves remarkable performance, even better than that on masked patches, where the latter is the standard format in the existing MIM methods. Secondly, the optimal mask ratio positively correlates to the model size. That is to say, the smaller the model, the lower the mask ratio needs to be. Driven by these two discoveries, our simple and concise approach CAE v2 achieves superior performance on a series of downstream tasks. For example, a vanilla ViT-Large model achieves 81.7% and 86.7% top-1 accuracy on linear probing and fine-tuning on ImageNet-1K, and 55.9% mIoU on semantic segmentation on ADE20K with the pre-training for 300 epochs. We hope our findings can be helpful guidelines for the pre-training in the MIM area, especially for the small-scale models.
Automated data augmentation, which aims at engineering augmentation policy automatically, recently draw a growing research interest. Many previous auto-augmentation methods utilized a Density Matching strategy by evaluating policies in terms of the test-time augmentation performance. In this paper, we theoretically and empirically demonstrated the inconsistency between the train and validation set of small-scale medical image datasets, referred to as in-domain sampling bias. Next, we demonstrated that the in-domain sampling bias might cause the inefficiency of Density Matching. To address the problem, an improved augmentation search strategy, named Augmented Density Matching, was proposed by randomly sampling policies from a prior distribution for training. Moreover, an efficient automatical machine learning(AutoML) algorithm was proposed by unifying the search on data augmentation and neural architecture. Experimental results indicated that the proposed methods outperformed state-of-the-art approaches on MedMNIST, a pioneering benchmark designed for AutoML in medical image analysis.
Concept bottleneck models (CBMs) include a bottleneck of human-interpretable concepts providing explainability and intervention during inference by correcting the predicted, intermediate concepts. This makes CBMs attractive for high-stakes decision-making. In this paper, we take the quality assessment of fetal ultrasound scans as a real-life use case for CBM decision support in healthcare. For this case, simple binary concepts are not sufficiently reliable, as they are mapped directly from images of highly variable quality, for which variable model calibration might lead to unstable binarized concepts. Moreover, scalar concepts do not provide the intuitive spatial feedback requested by users. To address this, we design a hierarchical CBM imitating the sequential expert decision-making process of "seeing", "conceiving" and "concluding". Our model first passes through a layer of visual, segmentation-based concepts, and next a second layer of property concepts directly associated with the decision-making task. We note that experts can intervene on both the visual and property concepts during inference. Additionally, we increase the bottleneck capacity by considering task-relevant concept interaction. Our application of ultrasound scan quality assessment is challenging, as it relies on balancing the (often poor) image quality against an assessment of the visibility and geometric properties of standardized image content. Our validation shows that -- in contrast with previous CBM models -- our CBM models actually outperform equivalent concept-free models in terms of predictive performance. Moreover, we illustrate how interventions can further improve our performance over the state-of-the-art.
In this paper, we study the partial multi-label (PML) image classification problem, where each image is annotated with a candidate label set consists of multiple relevant labels and other noisy labels. Existing PML methods typically design a disambiguation strategy to filter out noisy labels by utilizing prior knowledge with extra assumptions, which unfortunately is unavailable in many real tasks. Furthermore, because the objective function for disambiguation is usually elaborately designed on the whole training set, it can be hardly optimized in a deep model with SGD on mini-batches. In this paper, for the first time we propose a deep model for PML to enhance the representation and discrimination ability. On one hand, we propose a novel curriculum based disambiguation strategy to progressively identify ground-truth labels by incorporating the varied difficulties of different classes. On the other hand, a consistency regularization is introduced for model retraining to balance fitting identified easy labels and exploiting potential relevant labels. Extensive experimental results on the commonly used benchmark datasets show the proposed method significantly outperforms the SOTA methods.
One classical approach to regularize color is to tream them as two dimensional surfaces embedded in a five dimensional spatial-chromatic space. In this case, a natural regularization term arises as the image surface area. Choosing the chromatic coordinates as dominating over the spatial ones, the image spatial coordinates could be thought of as a paramterization of the image surface manifold in a three dimensional color space. Minimizing the area of the image manifold leads to the Beltrami flow or mean curvature flow of the image surface in the 3D color space, while minimizing the elastica of the image surface yields an additional interesting regularization. Recently, the authors proposed a color elastica model, which minimizes both the surface area and elastica of the image manifold. In this paper, we propose to modify the color elastica and introduce two new models for color image regularization. The revised measures are motivated by the relations between the color elastica model, Euler's elastica model and the total variation model for gray level images. Compared to our previous color elastica model, the new models are direct extensions of Euler's elastica model to color images. The proposed models are nonlinear and challenging to minimize. To overcome this difficulty, two operator-splitting methods are suggested. Specifically, nonlinearities are decoupled by introducing new vector- and matrix-valued variables. Then, the minimization problems are converted to solving initial value problems which are time-discretized by operator splitting. Each subproblem, after splitting either, has a closed-form solution or can be solved efficiently. The effectiveness and advantages of the proposed models are demonstrated by comprehensive experiments. The benefits of incorporating the elastica of the image surface as regularization terms compared to common alternatives are empirically validated.