Rectified Flow offers a simple and effective approach to high-quality generative modeling by learning a velocity field. However, we identify a limitation in directly modeling the velocity with an unconstrained neural network: the learned velocity often fails to satisfy certain boundary conditions, leading to inaccurate velocity field estimations that deviate from the desired ODE. This issue is particularly critical during stochastic sampling at inference, as the score function's errors are amplified near the boundary. To mitigate this, we propose a Boundary-enforced Rectified Flow Model (Boundary RF Model), in which we enforce boundary conditions with a minimal code modification. Boundary RF Model improves performance over vanilla RF model, demonstrating 8.01% improvement in FID score on ImageNet using ODE sampling and 8.98% improvement using SDE sampling.

Text-to-image generative models have recently attracted considerable interest, enabling the synthesis of high-quality images from textual prompts. However, these models often lack the capability to generate specific subjects from given reference images or to synthesize novel renditions under varying conditions. Methods like DreamBooth and Subject-driven Text-to-Image (SuTI) have made significant progress in this area. Yet, both approaches primarily focus on enhancing similarity to reference images and require expensive setups, often overlooking the need for efficient training and avoiding overfitting to the reference images. In this work, we present the $\lambda$-Harmonic reward function, which provides a reliable reward signal and enables early stopping for faster training and effective regularization. By combining the Bradley-Terry preference model, the $\lambda$-Harmonic reward function also provides preference labels for subject-driven generation tasks. We propose Reward Preference Optimization (RPO), which offers a simpler setup (requiring only 3\% of the negative samples used by DreamBooth) and fewer gradient steps for fine-tuning. Unlike most existing methods, our approach does not require training a text encoder or optimizing text embeddings and achieves text-image alignment by fine-tuning only the U-Net component. Empirically, $\lambda$-Harmonic proves to be a reliable approach for model selection in subject-driven generation tasks. Based on preference labels and early stopping validation from the $\lambda$-Harmonic reward function, our algorithm achieves a state-of-the-art CLIP-I score of 0.833 and a CLIP-T score of 0.314 on DreamBench.

In recent years, transformer-based models have dominated panoptic segmentation, thanks to their strong modeling capabilities and their unified representation for both semantic and instance classes as global binary masks. In this paper, we revisit pure convolution model and propose a novel panoptic architecture named MaskConver. MaskConver proposes to fully unify things and stuff representation by predicting their centers. To that extent, it creates a lightweight class embedding module that can break the ties when multiple centers co-exist in the same location. Furthermore, our study shows that the decoder design is critical in ensuring that the model has sufficient context for accurate detection and segmentation. We introduce a powerful ConvNeXt-UNet decoder that closes the performance gap between convolution- and transformer-based models. With ResNet50 backbone, our MaskConver achieves 53.6% PQ on the COCO panoptic val set, outperforming the modern convolution-based model, Panoptic FCN, by 9.3% as well as transformer-based models such as Mask2Former (+1.7% PQ) and kMaX-DeepLab (+0.6% PQ). Additionally, MaskConver with a MobileNet backbone reaches 37.2% PQ, improving over Panoptic-DeepLab by +6.4% under the same FLOPs/latency constraints. A further optimized version of MaskConver achieves 29.7% PQ, while running in real-time on mobile devices. The code and model weights will be publicly available.1

Sparsely activated Mixture-of-Experts (MoE) is becoming a promising paradigm for multi-task learning (MTL). Instead of compressing multiple tasks’ knowledge into a single model, MoE separates the parameter space and only utilizes the relevant model pieces given task type and its input, which provides stabilized MTL training and ultra-efficient inference. However, current MoE approaches adopt a fixed network capacity (e.g., two experts in usual) for all tasks. It potentially results in the over-fitting of simple tasks or the under-fitting of challenging scenarios, especially when tasks are significantly distinctive in their complexity. In this paper, we propose an adaptive MoE framework for multi-task vision recognition, dubbed AdaMV-MoE. Based on the training dynamics, it automatically determines the number of activated experts for each task, avoiding the laborious manual tuning of optimal model size. To validate our proposal, we benchmark it on ImageNet classification and COCO object detection & instance segmentation which are notoriously difficult to learn in concert, due to their discrepancy. Extensive experiments across a variety of vision transformers demonstrate a superior performance of AdaMV-MoE, compared to MTL with a shared backbone and the recent state-of-the-art (SoTA) MTL MoE approach. Codes are available online: https://github.com/google-research/google-research/tree/master/moe_mtl.

Recent studies have shown that there exhibits significantly imbalanced medical resource allocation across public hospitals. Patients, regardless of their diseases, tend to choose hospitals and physicians with a better reputation, which often overloads major hospitals while leaving others underutilized. Guiding patients to hospitals that can serve their treatment needs both timely and with good quality can make the best use of precious medical resources. Unfortunately, it remains one of the major challenges both for research and in practice. In this article, we propose a novel diversity-enhanced hierarchical physician recommendation approach to address this issue. We adopt matrix factorization to estimate physician competency and exploit implicit similarity relationships to improve the competency estimation of physicians that we are of little information of. We then balance the patient preference and physician diversity using two novel heuristic algorithms. We evaluate our proposed approach and compare it with the state of the art. Experiments show that our approach significantly improves both accuracy and recommendation diversity over existing approaches.

This paper introduces a video dataset of spatio-temporally localized Atomic Visual Actions (AVA). The AVA dataset densely annotates 80 atomic visual actions in 437 15-minute video clips, where actions are localized in space and time, resulting in 1.59M action labels with multiple labels per person occurring frequently. The key characteristics of our dataset are: (1) the definition of atomic visual actions, rather than composite actions; (2) precise spatio-temporal annotations with possibly multiple annotations for each person; (3) exhaustive annotation of these atomic actions over 15-minute video clips; (4) people temporally linked across consecutive segments; and (5) using movies to gather a varied set of action representations. This departs from existing datasets for spatio-temporal action recognition, which typically provide sparse annotations for composite actions in short video clips. AVA, with its realistic scene and action complexity, exposes the intrinsic difficulty of action recognition. To benchmark this, we present a novel approach for action localization that builds upon the current state-of-the-art methods, and demonstrates better performance on JHMDB and UCF101-24 categories. While setting a new state of the art on existing datasets, the overall results on AVA are low at 15.8% mAP, underscoring the need for developing new approaches for video understanding.

Similarity measure is an essential component in image registration. In this article, we propose a novel similarity measure for registration of two or more images. The proposed method is motivated by the fact that optimally registered images can be sparsified hierarchically in the gradient domain and frequency domain with the separation of sparse errors. One of the key advantages of the proposed similarity measure is its robustness in dealing with severe intensity distortions, which widely exist on medical images, remotely sensed images and natural photos due to differences of acquisition modalities or illumination conditions. Two efficient algorithms are proposed to solve the batch image registration and pair registration problems in a unified framework. We have validated our method on extensive and challenging data sets. The experimental results demonstrate the robustness, accuracy and efficiency of our method over nine traditional and state-of-the-art algorithms on synthetic images and a wide range of real-world applications.

Recently with the explosive growth of visual content on the Internet, large-scale image search has attracted intensive attention. It has been shown that mapping high-dimensional image descriptors to compact binary codes can lead to considerable efficiency gains in both storage and performing similarity computation of images. However, most existing methods still suffer from expensive training devoted to large-scale binary code learning. To address this issue, we propose a sub-selection based matrix manipulation algorithm, which can significantly reduce the computational cost of code learning. As case studies, we apply the sub-selection algorithm to several popular quantization techniques including cases using linear and nonlinear mappings. Crucially, we can justify the resulting sub-selective quantization by proving its theoretic properties. Extensive experiments are carried out on three image benchmarks with up to one million samples, corroborating the efficacy of the sub-selective quantization method in terms of image retrieval.

We present a neural network model — based on Convolutional Neural Networks, Recurrent Neural Networks and a novel attention mechanism — which achieves 84.2% accuracy on the challenging French Street Name Signs (FSNS) dataset, significantly outperforming the previous state of the art (Smith'16), which achieved 72.46%. Furthermore, our new method is much simpler and more general than the previous approach. To demonstrate the generality of our model, we show that it also performs well on an even more challenging dataset derived from Google Street View, in which the goal is to extract business names from store fronts. Finally, we study the speed/accuracy tradeoff that results from using CNN feature extractors of different depths. Surprisingly, we find that deeper is not always better (in terms of accuracy, as well as speed). Our resulting model is simple, accurate and fast, allowing it to be used at scale on a variety of challenging real-world text extraction problems.

In the past decades, Spectral Clustering (SC) has become one of the most effective clustering approaches. Although it has been widely used, one significant drawback of SC is its expensive computation cost. Many efforts have been devoted to accelerating SC algorithms and promising results have been achieved. However, most of the existing algorithms rely on the assumption that data can be stored in the computer memory. When data cannot fit in the memory, these algorithms will suffer severe performance degradations. In order to overcome this issue, we propose a novel sequential SC algorithm for tackling large-scale clustering with limited computational resources, \textit{e.g.}, memory. We begin with investigating an effective way of approximating the graph affinity matrix via leveraging a bipartite graph. Then we choose a smart graph construction and optimization strategy to avoid random access to data. These efforts lead to an efficient SC algorithm whose memory usage is independent of the number of input data points. Extensive experiments carried out on large datasets demonstrate that the proposed sequential SC algorithm is up to a thousand times faster than the state-of-the-arts.

Group sparse representation (GSR) exploits group structure in data and works well on many problems. However, the group structure must be manually given in advance. In many practical scenarios such as classification, samples are grouped according to their labels. Constructing a consistent group structure in such cases is not easy. The reasons are: 1) samples may be incorrectly labeled; and 2) label assigning in big data is time-consuming and expensive. In this paper, we propose and formulate a new problem, semi-supervised group sparse representation (SS-GSR) to support group sparse representation among both labeled and unlabeled data, while learning a more robust group structure, which can be further exploited to more effectively represent other unlabeled data. We develop a model to tackle the SS-GSR problem, based on the manifold assumption in subspace segmentation that samples in the same group lie close in feature space and span the same subspace. We also propose an alternating algorithm to solve the model. Finally, we validate the model via extensive experiments.

In this paper, we propose an algorithm for missing value recovery of visual data such as image or video. These missing values may result from the corruption in acquisition process, or user-specified unexpected outliers. This problem exists in wide range of applications. We use the nuclear norm (NN) regularization to enforce the global consistency of the image, while the total variation (TV) regularization is used to encourage the locally consistent in image intensity domain. This model can be applied in very challenging scenarios, where only very small amount of data is available. However, it is very difficult to efficiently solve these two regularizations simultaneously by convex programming due to its composite structure and non-smoothness. To this end, we propose an efficient proximal-splitting algorithm for joint NN/TV minimization. The proposed algorithm is theoretically guaranteed to achieve a convergence rate of O(1/N) for N iterations, which is much faster than O(1/√N) by the black-box first-order method for solving the non-smooth optimization problem. In our experiments, we demonstrate the superior performance of our algorithm on image completion compared with seven state-of-the-art algorithms.

In this paper, we address the problem of large-scale multi-view spectral clustering. In many real-world applications, data can be represented in various heterogeneous features or views. Different views often provide different aspects of information that are complementary to each other. Several previous methods of clustering have demonstrated that better accuracy can be achieved using integrated information of all the views than just using each view individually. One important class of such methods is multi-view spectral clustering, which is based on graph Laplacian. However, existing methods are not applicable to large-scale problem for their high computational complexity. To this end, we propose a novel large-scale multi-view spectral clustering approach based on the bipartite graph. Our method uses local manifold fusion to integrate heterogeneous features. To improve efficiency, we approximate the similarity graphs using bipartite graphs. Furthermore, we show that our method can be easily extended to handle the out-of-sample problem. Extensive experimental results on five benchmark datasets demonstrate the effectiveness and efficiency of the proposed method, where our method runs up to nearly 3000 times faster than the state-of-the-art methods.

The definition of the similarity measure is an essential component in image registration. In this paper, we propose a novel similarity measure for registration of two or more images. The proposed method is motivated by that the optimally registered images can be deeply sparsified in the gradient domain and frequency domain, with the separation of a sparse tensor of errors. One of the key advantages of the proposed similarity measure is its robustness to severe intensity distortions, which widely exist on medical images, remotely sensed images and natural photos due to the difference of acquisition modalities or illumination conditions. Two efficient algorithms are proposed to solve the batch image registration and pair registration problems in a unified framework. We validate our method on extensive challenging datasets. The experimental results demonstrate the robustness, accuracy and efficiency of our method over 9 traditional and state-of-the-art algorithms on synthetic images and a wide range of real-world applications.

In this paper, we propose a novel method for image fusion with a high-resolution panchromatic image and a low-resolution multispectral (Ms) image at the same geographical location. The fusion is formulated as a convex optimization problem which minimizes a linear combination of a least-squares fitting term and a dynamic gradient sparsity regularizer. The former is to preserve accurate spectral information of the Ms image, while the latter is to keep sharp edges of the high-resolution panchromatic image. We further propose to simultaneously register the two images during the fusing process, which is naturally achieved by virtue of the dynamic gradient sparsity property. An efficient algorithm is then devised to solve the optimization problem, accomplishing a linear computational complexity in the size of the output image in each iteration. We compare our method against six state-of-the-art image fusion methods on Ms image data sets from four satellites. Extensive experimental results demonstrate that the proposed method substantially outperforms the others in terms of both spatial and spectral qualities. We also show that our method can provide high-quality products from coarsely registered real-world IKONOS data sets. Finally, a MATLAB implementation is provided to facilitate future research.

Recently with the explosive growth of visual content on the Internet, large-scale image search has attracted intensive attention. It has been shown that mapping highdimensional image descriptors to compact binary codes can lead to considerable efficiency gains in both storage and similarity computation of images. However, most existing methods still suffer from expensive training devoted to large-scale binary code learning. To address this issue, we propose a sub-selection based matrix manipulation algorithm which can significantly reduce the computational cost of code learning. As case studies, we apply the sub-selection algorithm to two popular quantization techniques PCA Quantization (PCAQ) and Iterative Quantization (ITQ). Crucially, we can justify the resulting sub-selective quantization by proving its theoretic properties. Extensive experiments are carried out on three image benchmarks with up to one million samples, corroborating the efficacy of the sub-selective quantization method in terms of image retrieval.

In this paper, we propose a novel method for image fusion from a high resolution panchromatic image and a low resolution multispectral image at the same geographical location. Different from previous methods, we do not make any assumption about the upsampled multispectral image, but only assume that the fused image after downsampling should be close to the original multispectral image. This is a severely ill-posed problem and a dynamic gradient sparsity penalty is thus proposed for regularization. Incorporating the intra- correlations of different bands, this penalty can effectively exploit the prior information (e.g. sharp boundaries) from the panchromatic image. A new convex optimization algorithm is proposed to efficiently solve this problem. Extensive experiments on four multispectral datasets demonstrate that the proposed method significantly outperforms the state-of-the-arts in terms of both spatial and spectral qualities.

In this paper, we present an efficient and robust image representation method that can handle misalignment, occlusion and big noises with lower computational cost. It is motivated by the sub-selection technique, which uses partial observations to efficiently approximate the original high dimensional problems. While it is very efficient, their method can not handle many real problems in practical applications, such as misalignment, occlusion and big noises. To this end, we propose a robust sub-representation method, which can effectively handle these problems with an efficient scheme. While its performance guarantee was theoretically proved, numerous experiments on practical applications have further demonstrated that the proposed method can lead to significant performance improvement in terms of speed and accuracy.