Xiao Liu

Jiazheng Xu, Xiao Liu, Yuchen Wu et al. · tsinghua

We present a comprehensive solution to learn and improve text-to-image models from human preference feedback. To begin with, we build ImageReward -- the first general-purpose text-to-image human preference reward model -- to effectively encode human preferences. Its training is based on our systematic annotation pipeline including rating and ranking, which collects 137k expert comparisons to date. In human evaluation, ImageReward outperforms existing scoring models and metrics, making it a promising automatic metric for evaluating text-to-image synthesis. On top of it, we propose Reward Feedback Learning (ReFL), a direct tuning algorithm to optimize diffusion models against a scorer. Both automatic and human evaluation support ReFL's advantages over compared methods. All code and datasets are provided at \url{https://github.com/THUDM/ImageReward}.

Pedro Sanchez, Antanas Kascenas, Xiao Liu et al.

Reducing the requirement for densely annotated masks in medical image segmentation is important due to cost constraints. In this paper, we consider the problem of inferring pixel-level predictions of brain lesions by only using image-level labels for training. By leveraging recent advances in generative diffusion probabilistic models (DPM), we synthesize counterfactuals of "How would a patient appear if X pathology was not present?". The difference image between the observed patient state and the healthy counterfactual can be used for inferring the location of pathology. We generate counterfactuals that correspond to the minimal change of the input such that it is transformed to healthy domain. This requires training with healthy and unhealthy data in DPMs. We improve on previous counterfactual DPMs by manipulating the generation process with implicit guidance along with attention conditioning instead of using classifiers. Code is available at https://github.com/vios-s/Diff-SCM.

Xiao Liu, Spyridon Thermos, Pedro Sanchez et al.

Training medical image segmentation models usually requires a large amount of labeled data. By contrast, humans can quickly learn to accurately recognise anatomy of interest from medical (e.g. MRI and CT) images with some limited guidance. Such recognition ability can easily generalise to new images from different clinical centres. This rapid and generalisable learning ability is mostly due to the compositional structure of image patterns in the human brain, which is less incorporated in medical image segmentation. In this paper, we model the compositional components (i.e. patterns) of human anatomy as learnable von-Mises-Fisher (vMF) kernels, which are robust to images collected from different domains (e.g. clinical centres). The image features can be decomposed to (or composed by) the components with the composing operations, i.e. the vMF likelihoods. The vMF likelihoods tell how likely each anatomical part is at each position of the image. Hence, the segmentation mask can be predicted based on the vMF likelihoods. Moreover, with a reconstruction module, unlabeled data can also be used to learn the vMF kernels and likelihoods by recombining them to reconstruct the input image. Extensive experiments show that the proposed vMFNet achieves improved generalisation performance on two benchmarks, especially when annotations are limited. Code is publicly available at: https://github.com/vios-s/vMFNet.

Ye Yuan, Xiao Liu, Wondimu Dikubab et al.

Handwritten mathematical expression recognition (HMER) is a challenging task that has many potential applications. Recent methods for HMER have achieved outstanding performance with an encoder-decoder architecture. However, these methods adhere to the paradigm that the prediction is made "from one character to another", which inevitably yields prediction errors due to the complicated structures of mathematical expressions or crabbed handwritings. In this paper, we propose a simple and efficient method for HMER, which is the first to incorporate syntax information into an encoder-decoder network. Specifically, we present a set of grammar rules for converting the LaTeX markup sequence of each expression into a parsing tree; then, we model the markup sequence prediction as a tree traverse process with a deep neural network. In this way, the proposed method can effectively describe the syntax context of expressions, alleviating the structure prediction errors of HMER. Experiments on three benchmark datasets demonstrate that our method achieves better recognition performance than prior arts. To further validate the effectiveness of our method, we create a large-scale dataset consisting of 100k handwritten mathematical expression images acquired from ten thousand writers. The source code, new dataset, and pre-trained models of this work will be publicly available.

Xin Lin, Chao Ren, Xiao Liu et al.

Deep learning methods have shown remarkable performance in image denoising, particularly when trained on large-scale paired datasets. However, acquiring such paired datasets for real-world scenarios poses a significant challenge. Although unsupervised approaches based on generative adversarial networks offer a promising solution for denoising without paired datasets, they are difficult in surpassing the performance limitations of conventional GAN-based unsupervised frameworks without significantly modifying existing structures or increasing the computational complexity of denoisers. To address this problem, we propose a SC strategy for multiple denoisers. This strategy can achieve significant performance improvement without increasing the inference complexity of the GAN-based denoising framework. Its basic idea is to iteratively replace the previous less powerful denoiser in the filter-guided noise extraction module with the current powerful denoiser. This process generates better synthetic clean-noisy image pairs, leading to a more powerful denoiser for the next iteration. This baseline ensures the stability and effectiveness of the training network. The experimental results demonstrate the superiority of our method over state-of-the-art unsupervised methods.

Lijun Zhang, Xiao Liu, Hui Guan

Multi-task learning (MTL) jointly learns a set of tasks by sharing parameters among tasks. It is a promising approach for reducing storage costs while improving task accuracy for many computer vision tasks. The effective adoption of MTL faces two main challenges. The first challenge is to determine what parameters to share across tasks to optimize for both memory efficiency and task accuracy. The second challenge is to automatically apply MTL algorithms to an arbitrary CNN backbone without requiring time-consuming manual re-implementation and significant domain expertise. This paper addresses the challenges by developing the first programming framework AutoMTL that automates efficient MTL model development for vision tasks. AutoMTL takes as inputs an arbitrary backbone convolutional neural network (CNN) and a set of tasks to learn, and automatically produces a multi-task model that achieves high accuracy and small memory footprint simultaneously. Experiments on three popular MTL benchmarks (CityScapes, NYUv2, Tiny-Taskonomy) demonstrate the effectiveness of AutoMTL over state-of-the-art approaches as well as the generalizability of AutoMTL across CNNs. AutoMTL is open-sourced and available at https://github.com/zhanglijun95/AutoMTL.

Baichuan Huang, Hongwei Yi, Can Huang et al.

The present Multi-view stereo (MVS) methods with supervised learning-based networks have an impressive performance comparing with traditional MVS methods. However, the ground-truth depth maps for training are hard to be obtained and are within limited kinds of scenarios. In this paper, we propose a novel unsupervised multi-metric MVS network, named M^3VSNet, for dense point cloud reconstruction without any supervision. To improve the robustness and completeness of point cloud reconstruction, we propose a novel multi-metric loss function that combines pixel-wise and feature-wise loss function to learn the inherent constraints from different perspectives of matching correspondences. Besides, we also incorporate the normal-depth consistency in the 3D point cloud format to improve the accuracy and continuity of the estimated depth maps. Experimental results show that M3VSNet establishes the state-of-the-arts unsupervised method and achieves comparable performance with previous supervised MVSNet on the DTU dataset and demonstrates the powerful generalization ability on the Tanks and Temples benchmark with effective improvement. Our code is available at https://github.com/whubaichuan/M3VSNet

Xiao Liu, Spyridon Thermos, Gabriele Valvano et al.

A recent spate of state-of-the-art semi- and un-supervised solutions disentangle and encode image "content" into a spatial tensor and image appearance or "style" into a vector, to achieve good performance in spatially equivariant tasks (e.g. image-to-image translation). To achieve this, they employ different model design, learning objective, and data biases. While considerable effort has been made to measure disentanglement in vector representations, and assess its impact on task performance, such analysis for (spatial) content - style disentanglement is lacking. In this paper, we conduct an empirical study to investigate the role of different biases in content-style disentanglement settings and unveil the relationship between the degree of disentanglement and task performance. In particular, we consider the setting where we: (i) identify key design choices and learning constraints for three popular content-style disentanglement models; (ii) relax or remove such constraints in an ablation fashion; and (iii) use two metrics to measure the degree of disentanglement and assess its effect on each task performance. Our experiments reveal that there is a "sweet spot" between disentanglement, task performance and - surprisingly - content interpretability, suggesting that blindly forcing for higher disentanglement can hurt model performance and content factors semanticness. Our findings, as well as the used task-independent metrics, can be used to guide the design and selection of new models for tasks where content-style representations are useful.

Chaohao Xie, Shaohui Liu, Chao Li et al.

Most convolutional network (CNN)-based inpainting methods adopt standard convolution to indistinguishably treat valid pixels and holes, making them limited in handling irregular holes and more likely to generate inpainting results with color discrepancy and blurriness. Partial convolution has been suggested to address this issue, but it adopts handcrafted feature re-normalization, and only considers forward mask-updating. In this paper, we present a learnable attention map module for learning feature renormalization and mask-updating in an end-to-end manner, which is effective in adapting to irregular holes and propagation of convolution layers. Furthermore, learnable reverse attention maps are introduced to allow the decoder of U-Net to concentrate on filling in irregular holes instead of reconstructing both holes and known regions, resulting in our learnable bidirectional attention maps. Qualitative and quantitative experiments show that our method performs favorably against state-of-the-arts in generating sharper, more coherent and visually plausible inpainting results. The source code and pre-trained models will be available.

Chao Li, Dongliang He, Xiao Liu et al.

Recently, image super-resolution has been widely studied and achieved significant progress by leveraging the power of deep convolutional neural networks. However, there has been limited advancement in video super-resolution (VSR) due to the complex temporal patterns in videos. In this paper, we investigate how to adapt state-of-the-art methods of image super-resolution for video super-resolution. The proposed adapting method is straightforward. The information among successive frames is well exploited, while the overhead on the original image super-resolution method is negligible. Furthermore, we propose a learning-based method to ensemble the outputs from multiple super-resolution models. Our methods show superior performance and rank second in the NTIRE2019 Video Super-Resolution Challenge Track 1.

Jian Wang, Feng Zhou, Shilei Wen et al.

The modern image search system requires semantic understanding of image, and a key yet under-addressed problem is to learn a good metric for measuring the similarity between images. While deep metric learning has yielded impressive performance gains by extracting high level abstractions from image data, a proper objective loss function becomes the central issue to boost the performance. In this paper, we propose a novel angular loss, which takes angle relationship into account, for learning better similarity metric. Whereas previous metric learning methods focus on optimizing the similarity (contrastive loss) or relative similarity (triplet loss) of image pairs, our proposed method aims at constraining the angle at the negative point of triplet triangles. Several favorable properties are observed when compared with conventional methods. First, scale invariance is introduced, improving the robustness of objective against feature variance. Second, a third-order geometric constraint is inherently imposed, capturing additional local structure of triplet triangles than contrastive loss or triplet loss. Third, better convergence has been demonstrated by experiments on three publicly available datasets.