In real acoustic environment, speech enhancement is an arduous task to improve the quality and intelligibility of speech interfered by background noise and reverberation. Over the past years, deep learning has shown great potential on speech enhancement. In this paper, we propose a novel real-time framework called DBNet which is a dual-branch structure with alternate interconnection. Each branch incorporates an encoder-decoder architecture with skip connections. The two branches are responsible for spectrum and waveform modeling, respectively. A bridge layer is adopted to exchange information between the two branches. Systematic evaluation and comparison show that the proposed system substantially outperforms related algorithms under very challenging environments. And in INTERSPEECH 2021 Deep Noise Suppression (DNS) challenge, the proposed system ranks the top 8 in real-time track 1 in terms of the Mean Opinion Score (MOS) of the ITU-T P.835 framework.
Face clustering is a useful tool for applications like automatic face annotation and retrieval. The main challenge is that it is difficult to cluster images from the same identity with different face poses, occlusions, and image quality. Traditional clustering methods usually ignore the relationship between individual images and their neighbors which may contain useful context information. In this paper, we repurpose the well-known Transformer and introduce a Face Transformer for supervised face clustering. In Face Transformer, we decompose the face clustering into two steps: relation encoding and linkage predicting. Specifically, given a face image, a \textbf{relation encoder} module aggregates local context information from its neighbors and a \textbf{linkage predictor} module judges whether a pair of images belong to the same cluster or not. In the local linkage graph view, Face Transformer can generate more robust node and edge representations compared to existing methods. Experiments on both MS-Celeb-1M and DeepFashion show that our method achieves state-of-the-art performance, e.g., 91.12\% in pairwise F-score on MS-Celeb-1M.
Recently, significant progress has been made on semantic segmentation. However, the success of supervised semantic segmentation typically relies on a large amount of labelled data, which is time-consuming and costly to obtain. Inspired by the success of semi-supervised learning methods in image classification, here we propose a simple yet effective semi-supervised learning framework for semantic segmentation. We demonstrate that the devil is in the details: a set of simple design and training techniques can collectively improve the performance of semi-supervised semantic segmentation significantly. Previous works [3, 27] fail to employ strong augmentation in pseudo label learning efficiently, as the large distribution change caused by strong augmentation harms the batch normalisation statistics. We design a new batch normalisation, namely distribution-specific batch normalisation (DSBN) to address this problem and demonstrate the importance of strong augmentation for semantic segmentation. Moreover, we design a self correction loss which is effective in noise resistance. We conduct a series of ablation studies to show the effectiveness of each component. Our method achieves state-of-the-art results in the semi-supervised settings on the Cityscapes and Pascal VOC datasets.
The framework of dominant learned video compression methods is usually composed of motion prediction modules as well as motion vector and residual image compression modules, suffering from its complex structure and error propagation problem. Approaches have been proposed to reduce the complexity by replacing motion prediction modules with implicit flow networks. Error propagation aware training strategy is also proposed to alleviate incremental reconstruction errors from previously decoded frames. Although these methods have brought some improvement, little attention has been paid to the framework itself. Inspired by the success of learned image compression through simplifying the framework with a single deep neural network, it is natural to expect a better performance in video compression via a simple yet appropriate framework. Therefore, we propose a framework to directly compress raw-pixel frames (rather than residual images), where no extra motion prediction module is required. Instead, an entropy model is used to estimate the spatiotemporal redundancy in a latent space rather than pixel level, which significantly reduces the complexity of the framework. Specifically, the whole framework is a compression module, consisting of a unified auto-encoder which produces identically distributed latents for all frames, and a spatiotemporal entropy estimation model to minimize the entropy of these latents. Experiments showed that the proposed method outperforms state-of-the-art (SOTA) performance under the metric of multiscale structural similarity (MS-SSIM) and achieves competitive results under the metric of PSNR.
Noisy labels are very common in deep supervised learning. Although many studies tend to improve the robustness of deep training for noisy labels, rare works focus on theoretically explaining the training behaviors of learning with noisily labeled data, which is a fundamental principle in understanding its generalization. In this draft, we study its two phenomena, clean data first and phase transition, by explaining them from a theoretical viewpoint. Specifically, we first show that in the first epoch training, the examples with clean labels will be learned first. We then show that after the learning from clean data stage, continuously training model can achieve further improvement in testing error when the rate of corrupted class labels is smaller than a certain threshold; otherwise, extensively training could lead to an increasing testing error.
Features that are equivariant to a larger group of symmetries have been shown to be more discriminative and powerful in recent studies. However, higher-order equivariant features often come with an exponentially-growing computational cost. Furthermore, it remains relatively less explored how rotation-equivariant features can be leveraged to tackle 3D shape alignment tasks. While many past approaches have been based on either non-equivariant or invariant descriptors to align 3D shapes, we argue that such tasks may benefit greatly from an equivariant framework. In this paper, we propose an effective and practical SE(3) (3D translation and rotation) equivariant network for point cloud analysis that addresses both problems. First, we present SE(3) separable point convolution, a novel framework that breaks down the 6D convolution into two separable convolutional operators alternatively performed in the 3D Euclidean and SO(3) spaces. This significantly reduces the computational cost without compromising the performance. Second, we introduce an attention layer to effectively harness the expressiveness of the equivariant features. While jointly trained with the network, the attention layer implicitly derives the intrinsic local frame in the feature space and generates attention vectors that can be integrated into different alignment tasks. We evaluate our approach through extensive studies and visual interpretations. The empirical results demonstrate that our proposed model outperforms strong baselines in a variety of benchmarks
Temporal action proposal generation (TAPG) is a fundamental and challenging task in video understanding, especially in temporal action detection. Most previous works focus on capturing the local temporal context and can well locate simple action instances with clean frames and clear boundaries. However, they generally fail in complicated scenarios where interested actions involve irrelevant frames and background clutters, and the local temporal context becomes less effective. To deal with these problems, we present an augmented transformer with adaptive graph network (ATAG) to exploit both long-range and local temporal contexts for TAPG. Specifically, we enhance the vanilla transformer by equipping a snippet actionness loss and a front block, dubbed augmented transformer, and it improves the abilities of capturing long-range dependencies and learning robust feature for noisy action instances.Moreover, an adaptive graph convolutional network (GCN) is proposed to build local temporal context by mining the position information and difference between adjacent features. The features from the two modules carry rich semantic information of the video, and are fused for effective sequential proposal generation. Extensive experiments are conducted on two challenging datasets, THUMOS14 and ActivityNet1.3, and the results demonstrate that our method outperforms state-of-the-art TAPG methods. Our code will be released soon.
Face detector frequently confronts extreme scale variance challenge. The famous solutions are Multi-scale training, Data-anchor-sampling and Random crop strategy. In this paper, we indicate 2 significant elements to resolve extreme scale variance problem by investigating the difference among the previous solutions, including the fore-ground and back-ground information of an image and the scale information. However, current excellent solutions can only utilize the former information while neglecting to absorb the latter one effectively. In order to help the detector utilize the scale information efficiently, we analyze the relationship between the detector performance and the scale distribution of the training data. Based on this analysis, we propose a Selective Scale Enhancement (SSE) strategy which can assimilate these two information efficiently and simultaneously. Finally, our method achieves state-of-the-art detection performance on all common face detection benchmarks, including AFW, PASCAL face, FDDB and Wider Face datasets. Note that our result achieves six champions on the Wider Face dataset.
Deep learning has shown a great potential for speech separation, especially for speech and non-speech separation. However, it encounters permutation problem for multi-speaker separation where both target and interference are speech. Permutation Invariant training (PIT) was proposed to solve this problem by permuting the order of the multiple speakers. Another way is to use an anchor speech, a short speech of the target speaker, to model the speaker identity. In this paper, we propose a simple strategy to train a long short-term memory (LSTM) model to solve the permutation problem in speaker separation. Specifically, we insert a short speech of target speaker at the beginning of a mixture as guide information. So, the first appearing speaker is defined as the target. Due to the powerful capability on sequence modeling, LSTM can use its memory cells to track and separate target speech from interfering speech. Experimental results show that the proposed training strategy is effective for speaker separation.