Chao Zhang

Xuesong Chen, Shaoshuai Shi, Chao Zhang et al.

3D multi-object tracking (MOT) is vital for many applications including autonomous driving vehicles and service robots. With the commonly used tracking-by-detection paradigm, 3D MOT has made important progress in recent years. However, these methods only use the detection boxes of the current frame to obtain trajectory-box association results, which makes it impossible for the tracker to recover objects missed by the detector. In this paper, we present TrajectoryFormer, a novel point-cloud-based 3D MOT framework. To recover the missed object by detector, we generates multiple trajectory hypotheses with hybrid candidate boxes, including temporally predicted boxes and current-frame detection boxes, for trajectory-box association. The predicted boxes can propagate object's history trajectory information to the current frame and thus the network can tolerate short-term miss detection of the tracked objects. We combine long-term object motion feature and short-term object appearance feature to create per-hypothesis feature embedding, which reduces the computational overhead for spatial-temporal encoding. Additionally, we introduce a Global-Local Interaction Module to conduct information interaction among all hypotheses and models their spatial relations, leading to accurate estimation of hypotheses. Our TrajectoryFormer achieves state-of-the-art performance on the Waymo 3D MOT benchmarks. Code is available at https://github.com/poodarchu/EFG .

Xingbo Dong, Wanyan Xu, Zhihui Miao et al.

Low-light image enhancement - a pervasive but challenging problem, plays a central role in enhancing the visibility of an image captured in a poor illumination environment. Due to the fact that not all photons can pass the Bayer-Filter on the sensor of the color camera, in this work, we first present a De-Bayer-Filter simulator based on deep neural networks to generate a monochrome raw image from the colored raw image. Next, a fully convolutional network is proposed to achieve the low-light image enhancement by fusing colored raw data with synthesized monochrome raw data. Channel-wise attention is also introduced to the fusion process to establish a complementary interaction between features from colored and monochrome raw images. To train the convolutional networks, we propose a dataset with monochrome and color raw pairs named Mono-Colored Raw paired dataset (MCR) collected by using a monochrome camera without Bayer-Filter and a color camera with Bayer-Filter. The proposed pipeline take advantages of the fusion of the virtual monochrome and the color raw images and our extensive experiments indicate that significant improvement can be achieved by leveraging raw sensor data and data-driven learning.

Zhaoyang Huang, Xiaoyu Shi, Chao Zhang et al.

This paper introduces a novel transformer-based network architecture, FlowFormer, along with the Masked Cost Volume AutoEncoding (MCVA) for pretraining it to tackle the problem of optical flow estimation. FlowFormer tokenizes the 4D cost-volume built from the source-target image pair and iteratively refines flow estimation with a cost-volume encoder-decoder architecture. The cost-volume encoder derives a cost memory with alternate-group transformer~(AGT) layers in a latent space and the decoder recurrently decodes flow from the cost memory with dynamic positional cost queries. On the Sintel benchmark, FlowFormer architecture achieves 1.16 and 2.09 average end-point-error~(AEPE) on the clean and final pass, a 16.5\% and 15.5\% error reduction from the GMA~(1.388 and 2.47). MCVA enhances FlowFormer by pretraining the cost-volume encoder with a masked autoencoding scheme, which further unleashes the capability of FlowFormer with unlabeled data. This is especially critical in optical flow estimation because ground truth flows are more expensive to acquire than labels in other vision tasks. MCVA improves FlowFormer all-sided and FlowFormer+MCVA ranks 1st among all published methods on both Sintel and KITTI-2015 benchmarks and achieves the best generalization performance. Specifically, FlowFormer+MCVA achieves 1.07 and 1.94 AEPE on the Sintel benchmark, leading to 7.76\% and 7.18\% error reductions from FlowFormer.

Chunzhi Gu, Jun Yu, Chao Zhang

Generative model-based motion prediction techniques have recently realized predicting controlled human motions, such as predicting multiple upper human body motions with similar lower-body motions. However, to achieve this, the state-of-the-art methods require either subsequently learning mapping functions to seek similar motions or training the model repetitively to enable control over the desired portion of body. In this paper, we propose a novel framework to learn disentangled representations for controllable human motion prediction. Our network involves a conditional variational auto-encoder (CVAE) architecture to model full-body human motion, and an extra CVAE path to learn only the corresponding partial-body (e.g., lower-body) motion. Specifically, the inductive bias imposed by the extra CVAE path encourages two latent variables in two paths to respectively govern separate representations for each partial-body motion. With a single training, our model is able to provide two types of controls for the generated human motions: (i) strictly controlling one portion of human body and (ii) adaptively controlling the other portion, by sampling from a pair of latent spaces. Additionally, we extend and adapt a sampling strategy to our trained model to diversify the controllable predictions. Our framework also potentially allows new forms of control by flexibly customizing the input for the extra CVAE path. Extensive experimental results and ablation studies demonstrate that our approach is capable of predicting state-of-the-art controllable human motions both qualitatively and quantitatively.

Chunzhi Gu, Chao Zhang, Shigeru Kuriyama

The task of action-driven human motion prediction aims to forecast future human motion based on the observed sequence while respecting the given action label. It requires modeling not only the stochasticity within human motion but the smooth yet realistic transition between multiple action labels. However, the fact that most datasets do not contain such transition data complicates this task. Existing work tackles this issue by learning a smoothness prior to simply promote smooth transitions, yet doing so can result in unnatural transitions especially when the history and predicted motions differ significantly in orientations. In this paper, we argue that valid human motion transitions should incorporate realistic leg movements to handle orientation changes, and cast it as an action-conditioned in-betweening (ACB) learning task to encourage transition naturalness. Because modeling all possible transitions is virtually unreasonable, our ACB is only performed on very few selected action classes with active gait motions, such as Walk or Run. Specifically, we follow a two-stage forecasting strategy by first employing the motion diffusion model to generate the target motion with a specified future action, and then producing the in-betweening to smoothly connect the observation and prediction to eventually address motion prediction. Our method is completely free from the labeled motion transition data during training. To show the robustness of our approach, we generalize our trained in-betweening learning model on one dataset to two unseen large-scale motion datasets to produce natural transitions. Extensive experimental evaluations on three benchmark datasets demonstrate that our method yields the state-of-the-art performance in terms of visual quality, prediction accuracy, and action faithfulness.

Katsuya Hotta, Chao Zhang, Yoshihiro Hagihara et al.

Unsupervised anomaly localization, which plays a critical role in industrial manufacturing, aims to identify anomalous regions that deviate from normal sample patterns. Most recent methods perform feature matching or reconstruction for the target sample with pre-trained deep neural networks. However, they still struggle to address challenging anomalies because the deep embeddings stored in the memory bank can be less powerful and informative. More specifically, prior methods often overly rely on the finite resources stored in the memory bank, which leads to low robustness to unseen targets. In this paper, we propose a novel subspace-guided feature reconstruction framework to pursue adaptive feature approximation for anomaly localization. It first learns to construct low-dimensional subspaces from the given nominal samples, and then learns to reconstruct the given deep target embedding by linearly combining the subspace basis vectors using the self-expressive model. Our core is that, despite the limited resources in the memory bank, the out-of-bank features can be alternatively ``mimicked'' under the self-expressive mechanism to adaptively model the target. Eventually, the poorly reconstructed feature dimensions indicate anomalies for localization. Moreover, we propose a sampling method that leverages the sparsity of subspaces and allows the feature reconstruction to depend only on a small resource subset, which contributes to less memory overhead. Extensive experiments on three industrial benchmark datasets demonstrate that our approach generally achieves state-of-the-art anomaly localization performance.

Qiang Zheng, Yafei Qi, Chen Wang et al.

In the domain of point cloud analysis, despite the significant capabilities of Graph Neural Networks (GNNs) in managing complex 3D datasets, existing approaches encounter challenges like high computational costs and scalability issues with extensive scenarios. These limitations restrict the practical deployment of GNNs, notably in resource-constrained environments. To address these issues, this study introduce <b>Point<\b> <b>Vi<\b>sion <b>G<\b>NN (PointViG), an efficient framework for point cloud analysis. PointViG incorporates a lightweight graph convolutional module to efficiently aggregate local features and mitigate over-smoothing. For large-scale point cloud scenes, we propose an adaptive dilated graph convolution technique that searches for sparse neighboring nodes within a dilated neighborhood based on semantic correlation, thereby expanding the receptive field and ensuring computational efficiency. Experiments demonstrate that PointViG achieves performance comparable to state-of-the-art models while balancing performance and complexity. On the ModelNet40 classification task, PointViG achieved 94.3% accuracy with 1.5M parameters. For the S3DIS segmentation task, it achieved an mIoU of 71.7% with 5.3M parameters. These results underscore the potential and efficiency of PointViG in point cloud analysis.

Qiang Zheng, Chao Zhang, Jian Sun

In recent years, point cloud analysis methods based on the Transformer architecture have made significant progress, particularly in the context of multimedia applications such as 3D modeling, virtual reality, and autonomous systems. However, the high computational resource demands of the Transformer architecture hinder its scalability, real-time processing capabilities, and deployment on mobile devices and other platforms with limited computational resources. This limitation remains a significant obstacle to its practical application in scenarios requiring on-device intelligence and multimedia processing. To address this challenge, we propose an efficient point cloud analysis architecture, \textbf{Point} \textbf{M}LP-\textbf{T}ransformer (PointMT). This study tackles the quadratic complexity of the self-attention mechanism by introducing a linear complexity local attention mechanism for effective feature aggregation. Additionally, to counter the Transformer's focus on token differences while neglecting channel differences, we introduce a parameter-free channel temperature adaptation mechanism that adaptively adjusts the attention weight distribution in each channel, enhancing the precision of feature aggregation. To improve the Transformer's slow convergence speed due to the limited scale of point cloud datasets, we propose an MLP-Transformer hybrid module, which significantly enhances the model's convergence speed. Furthermore, to boost the feature representation capability of point tokens, we refine the classification head, enabling point tokens to directly participate in prediction. Experimental results on multiple evaluation benchmarks demonstrate that PointMT achieves performance comparable to state-of-the-art methods while maintaining an optimal balance between performance and accuracy.

Yuhui Yuan, Rao Fu, Lang Huang et al.

We present a High-Resolution Transformer (HRFormer) that learns high-resolution representations for dense prediction tasks, in contrast to the original Vision Transformer that produces low-resolution representations and has high memory and computational cost. We take advantage of the multi-resolution parallel design introduced in high-resolution convolutional networks (HRNet), along with local-window self-attention that performs self-attention over small non-overlapping image windows, for improving the memory and computation efficiency. In addition, we introduce a convolution into the FFN to exchange information across the disconnected image windows. We demonstrate the effectiveness of the High-Resolution Transformer on both human pose estimation and semantic segmentation tasks, e.g., HRFormer outperforms Swin transformer by $1.3$ AP on COCO pose estimation with $50\%$ fewer parameters and $30\%$ fewer FLOPs. Code is available at: https://github.com/HRNet/HRFormer.

Kai Han, Jianyuan Guo, Chao Zhang et al.

How to learn a discriminative fine-grained representation is a key point in many computer vision applications, such as person re-identification, fine-grained classification, fine-grained image retrieval, etc. Most of the previous methods focus on learning metrics or ensemble to derive better global representation, which are usually lack of local information. Based on the considerations above, we propose a novel Attribute-Aware Attention Model ($A^3M$), which can learn local attribute representation and global category representation simultaneously in an end-to-end manner. The proposed model contains two attention models: attribute-guided attention module uses attribute information to help select category features in different regions, at the same time, category-guided attention module selects local features of different attributes with the help of category cues. Through this attribute-category reciprocal process, local and global features benefit from each other. Finally, the resulting feature contains more intrinsic information for image recognition instead of the noisy and irrelevant features. Extensive experiments conducted on Market-1501, CompCars, CUB-200-2011 and CARS196 demonstrate the effectiveness of our $A^3M$. Code is available at https://github.com/iamhankai/attribute-aware-attention.

Qiang Zheng, Chao Zhang, Jian Sun

The rapid advancement in point cloud processing technologies has significantly increased the demand for efficient and compact models that achieve high-accuracy classification. Knowledge distillation has emerged as a potent model compression technique. However, traditional KD often requires extensive computational resources for forward inference of large teacher models, thereby reducing training efficiency for student models and increasing resource demands. To address these challenges, we introduce an innovative offline recording strategy that avoids the simultaneous loading of both teacher and student models, thereby reducing hardware demands. This approach feeds a multitude of augmented samples into the teacher model, recording both the data augmentation parameters and the corresponding logit outputs. By applying shape-level augmentation operations such as random scaling and translation, while excluding point-level operations like random jittering, the size of the records is significantly reduced. Additionally, to mitigate the issue of small student model over-imitating the teacher model's outputs and converging to suboptimal solutions, we incorporate a negative-weight self-distillation strategy. Experimental results demonstrate that the proposed distillation strategy enables the student model to achieve performance comparable to state-of-the-art models while maintaining lower parameter count. This approach strikes an optimal balance between performance and complexity. This study highlights the potential of our method to optimize knowledge distillation for point cloud classification tasks, particularly in resource-constrained environments, providing a novel solution for efficient point cloud analysis.

Qiang Zheng, Chao Zhang, Jian Sun

Advances in self-supervised learning are essential for enhancing feature extraction and understanding in point cloud processing. This paper introduces PMT-MAE (Point MLP-Transformer Masked Autoencoder), a novel self-supervised learning framework for point cloud classification. PMT-MAE features a dual-branch architecture that integrates Transformer and MLP components to capture rich features. The Transformer branch leverages global self-attention for intricate feature interactions, while the parallel MLP branch processes tokens through shared fully connected layers, offering a complementary feature transformation pathway. A fusion mechanism then combines these features, enhancing the model's capacity to learn comprehensive 3D representations. Guided by the sophisticated teacher model Point-M2AE, PMT-MAE employs a distillation strategy that includes feature distillation during pre-training and logit distillation during fine-tuning, ensuring effective knowledge transfer. On the ModelNet40 classification task, achieving an accuracy of 93.6\% without employing voting strategy, PMT-MAE surpasses the baseline Point-MAE (93.2\%) and the teacher Point-M2AE (93.4\%), underscoring its ability to learn discriminative 3D point cloud representations. Additionally, this framework demonstrates high efficiency, requiring only 40 epochs for both pre-training and fine-tuning. PMT-MAE's effectiveness and efficiency render it well-suited for scenarios with limited computational resources, positioning it as a promising solution for practical point cloud analysis.

Qiang Zheng, Chao Zhang, Jian Sun

Point cloud classification plays a crucial role in the processing and analysis of data from 3D sensors such as LiDAR, which are commonly used in applications like autonomous vehicles, robotics, and environmental monitoring. However, traditional neural networks, which rely heavily on multiplication operations, often face challenges in terms of high computational costs and energy consumption. This study presents a novel family of efficient MLP-based architectures designed to improve the computational efficiency of point cloud classification tasks in sensor systems. The baseline model, Mul-MLP, utilizes conventional multiplication operations, while Add-MLP and Shift-MLP replace multiplications with addition and shift operations, respectively. These replacements leverage more sensor-friendly operations that can significantly reduce computational overhead, making them particularly suitable for resource-constrained sensor platforms. To further enhance performance, we propose SA-MLP, a hybrid architecture that alternates between shift and adder layers, preserving the network depth while optimizing computational efficiency. Unlike previous approaches such as ShiftAddNet, which increase the layer count and limit representational capacity by freezing shift weights, SA-MLP fully exploits the complementary advantages of shift and adder layers by employing distinct learning rates and optimizers. Experimental results show that Add-MLP and Shift-MLP achieve competitive performance compared to Mul-MLP, while SA-MLP surpasses the baseline, delivering results comparable to state-of-the-art MLP models in terms of both classification accuracy and computational efficiency. This work offers a promising, energy-efficient solution for sensor-driven applications requiring real-time point cloud classification, particularly in environments with limited computational resources.

Tao Ma, Chao Zhang, Min Lu et al.

Renal pathology, as the gold standard of kidney disease diagnosis, requires doctors to analyze a series of tissue slices stained by H&E staining and special staining like Masson, PASM, and PAS, respectively. These special staining methods are costly, time-consuming, and hard to standardize for wide use especially in primary hospitals. Advances of supervised learning methods have enabled the virtually conversion of H&E images into special staining images, but achieving pixel-to-pixel alignment for training remains challenging. In contrast, unsupervised learning methods regarding different stains as different style transfer domains can utilize unpaired data, but they ignore the spatial inter-domain correlations and thus decrease the trustworthiness of structural details for diagnosis. In this paper, we propose a novel virtual staining framework AGMDT to translate images into other domains by avoiding pixel-level alignment and meanwhile utilizing the correlations among adjacent tissue slices. We first build a high-quality multi-domain renal histological dataset where each specimen case comprises a series of slices stained in various ways. Based on it, the proposed framework AGMDT discovers patch-level aligned pairs across the serial slices of multi-domains through glomerulus detection and bipartite graph matching, and utilizes such correlations to supervise the end-to-end model for multi-domain staining transformation. Experimental results show that the proposed AGMDT achieves a good balance between the precise pixel-level alignment and unpaired domain transfer by exploiting correlations across multi-domain serial pathological slices, and outperforms the state-of-the-art methods in both quantitative measure and morphological details.

Hui Lin, Chao Zhang, Danfeng Hong et al.

Remote sensing data is often distributed across multiple institutions, and due to privacy concerns and data-sharing restrictions, leveraging large-scale datasets in a centralized training framework is challenging. Federated learning offers a promising solution by enabling collaborative model training across distributed data sources without requiring data centralization. However, current Vision-Language Models (VLMs), which typically contain billions of parameters, pose significant communication challenges for traditional federated learning approaches based on model parameter updates, as they would incur substantial communication costs. In this paper, we propose FedRSCLIP, the first federated learning framework designed for remote sensing image classification based on a VLM, specifically CLIP. FedRSCLIP addresses the challenges of data heterogeneity and large-scale model transmission in federated environments by introducing Prompt Learning, which optimizes only a small set of tunable parameters. The framework introduces a dual-prompt mechanism, comprising Shared Prompts for global knowledge sharing and Private Prompts for client-specific adaptation. To maintain semantic coherence between shared and private prompts, we propose the Dual Prompt Alignment Constraint to balance global consistency and local adaptability across diverse client distributions. Additionally, to enhance cross-modal representation learning, we introduce the Cross-Modal Feature Alignment Constraint to align multimodal features between text and image prompts. To validate the effectiveness of our proposed model, we construct a Fed-RSIC dataset based on three existing remote sensing image classification datasets, specifically designed to simulate various federated learning configurations. Experimental results demonstrate the effectiveness and superiority of FedRSCLIP in remote sensing image classification.

Ziyun Cui, Ziyang Zhang, Guangzhi Sun et al.

Advances in large language models raise the question of how alignment techniques will adapt as models become increasingly complex and humans will only be able to supervise them weakly. Weak-to-Strong mimics such a scenario where weak model supervision attempts to harness the full capabilities of a much stronger model. This work extends Weak-to-Strong to WeakS-to-Strong by exploring an ensemble of weak models which simulate the variability in human opinions. Confidence scores are estimated using a Bayesian approach to guide the WeakS-to-Strong generalization. Furthermore, we extend the application of WeakS-to-Strong from text classification tasks to text generation tasks where more advanced strategies are investigated for supervision. Moreover, direct preference optimization is applied to advance the student model's preference learning, beyond the basic learning framework of teacher forcing. Results demonstrate the effectiveness of the proposed approach for the reliability of a strong student model, showing potential for superalignment.

Zi Wang, Katsuya Hotta, Koichiro Kamide et al.

High-resolution 3D point clouds are highly effective for detecting subtle structural anomalies in industrial inspection. However, their dense and irregular nature imposes significant challenges, including high computational cost, sensitivity to spatial misalignment, and difficulty in capturing localized structural differences. This paper introduces a registration-based anomaly detection framework that combines multi-prototype alignment with cluster-wise discrepancy analysis to enable precise 3D anomaly localization. Specifically, each test sample is first registered to multiple normal prototypes to enable direct structural comparison. To evaluate anomalies at a local level, clustering is performed over the point cloud, and similarity is computed between features from the test sample and the prototypes within each cluster. Rather than selecting cluster centroids randomly, a keypoint-guided strategy is employed, where geometrically informative points are chosen as centroids. This ensures that clusters are centered on feature-rich regions, enabling more meaningful and stable distance-based comparisons. Extensive experiments on the Real3D-AD benchmark demonstrate that the proposed method achieves state-of-the-art performance in both object-level and point-level anomaly detection, even using only raw features.

Sadegh Nadimi, Vasileios Angelidakis, Sadaf Maramizonouz et al.

We introduce OCULAR, an innovative hardware and software solution for three-dimensional dynamic image analysis of fine particles. Current state-of-the art instruments for dynamic image analysis are largely limited to two-dimensional imaging. However, extensive literature has demonstrated that relying on a single two-dimensional projection for particle characterisation can lead to inaccuracies in many applications. Existing three-dimensional imaging technologies, such as computed tomography, laser scanning, and orthophotography, are limited to static objects. These methods are often not statistically representative and come with significant post-processing requirements, as well as the need for specialised imaging and computing resources. OCULAR addresses these challenges by providing a cost-effective solution for imaging continuous particle streams using a synchronised array of optical cameras. Particle shape characterisation is achieved through the reconstruction of their three-dimensional surfaces. This paper details the OCULAR methodology, evaluates its reproducibility, and compares its results against X-ray micro computed tomography, highlighting its potential for efficient and reliable particle analysis.

Katsuya Hotta, Takuya Akashi, Shogo Tokai et al.

Subspace clustering methods which embrace a self-expressive model that represents each data point as a linear combination of other data points in the dataset provide powerful unsupervised learning techniques. However, when dealing with large datasets, representation of each data point by referring to all data points via a dictionary suffers from high computational complexity. To alleviate this issue, we introduce a parallelizable multi-subset based self-expressive model (PMS) which represents each data point by combining multiple subsets, with each consisting of only a small proportion of the samples. The adoption of PMS in subspace clustering (PMSSC) leads to computational advantages because the optimization problems decomposed over each subset are small, and can be solved efficiently in parallel. Furthermore, PMSSC is able to combine multiple self-expressive coefficient vectors obtained from subsets, which contributes to an improvement in self-expressiveness. Extensive experiments on synthetic and real-world datasets show the efficiency and effectiveness of our approach in comparison to other methods.

Chunzhi Gu, Shuofeng Zhao, Chao Zhang

In this paper, we present a deep generative model based method to generate diverse human motion interpolation results. We resort to the Conditional Variational Auto-Encoder (CVAE) to learn human motion conditioned on a pair of given start and end motions, by leveraging the Recurrent Neural Network (RNN) structure for both the encoder and the decoder. Additionally, we introduce a regularization loss to further promote sample diversity. Once trained, our method is able to generate multiple plausible coherent motions by repetitively sampling from the learned latent space. Experiments on the publicly available dataset demonstrate the effectiveness of our method, in terms of sample plausibility and diversity.

Takumi Nakane, Haoran Xie, Chao Zhang

The free-form deformation model can represent a wide range of non-rigid deformations by manipulating a control point lattice over the image. However, due to a large number of parameters, it is challenging to fit the free-form deformation model directly to the deformed image for deformation estimation because of the complexity of the fitness landscape. In this paper, we cast the registration task as a multi-objective optimization problem (MOP) according to the fact that regions affected by each control point overlap with each other. Specifically, by partitioning the template image into several regions and measuring the similarity of each region independently, multiple objectives are built and deformation estimation can thus be realized by solving the MOP with off-the-shelf multi-objective evolutionary algorithms (MOEAs). In addition, a coarse-to-fine strategy is realized by image pyramid combined with control point mesh subdivision. Specifically, the optimized candidate solutions of the current image level are inherited by the next level, which increases the ability to deal with large deformation. Also, a post-processing procedure is proposed to generate a single output utilizing the Pareto optimal solutions. Comparative experiments on both synthetic and real-world images show the effectiveness and usefulness of our deformation estimation method.

Wei Wang, Chao Zhang, Xiaopei Wu

Accent recognition with deep learning framework is a similar work to deep speaker identification, they're both expected to give the input speech an identifiable representation. Compared with the individual-level features learned by speaker identification network, the deep accent recognition work throws a more challenging point that forging group-level accent features for speakers. In this paper, we borrow and improve the deep speaker identification framework to recognize accents, in detail, we adopt Convolutional Recurrent Neural Network as front-end encoder and integrate local features using Recurrent Neural Network to make an utterance-level accent representation. Novelly, to address overfitting, we simply add Connectionist Temporal Classification based speech recognition auxiliary task during training, and for ambiguous accent discrimination, we introduce some powerful discriminative loss functions in face recognition works to enhance the discriminative power of accent features. We show that our proposed network with discriminative training method (without data-augment) is significantly ahead of the baseline system on the accent classification track in the Accented English Speech Recognition Challenge 2020, where the loss function Circle-Loss has achieved the best discriminative optimization for accent representation.

Chunzhi Gu, Xuequan Lu, Chao Zhang

Color transfer, which plays a key role in image editing, has attracted noticeable attention recently. It has remained a challenge to date due to various issues such as time-consuming manual adjustments and prior segmentation issues. In this paper, we propose to model color transfer under a probability framework and cast it as a parameter estimation problem. In particular, we relate the transferred image with the example image under the Gaussian Mixture Model (GMM) and regard the transferred image color as the GMM centroids. We employ the Expectation-Maximization (EM) algorithm (E-step and M-step) for optimization. To better preserve gradient information, we introduce a Laplacian based regularization term to the objective function at the M-step which is solved by deriving a gradient descent algorithm. Given the input of a source image and an example image, our method is able to generate continuous color transfer results with increasing EM iterations. Various experiments show that our approach generally outperforms other competitive color transfer methods, both visually and quantitatively.

Dongbo Zhang, Zheng Fang, Xuequan Lu et al.

3D human segmentation has seen noticeable progress in re-cent years. It, however, still remains a challenge to date. In this paper, weintroduce a deep patch-based method for 3D human segmentation. Wefirst extract a local surface patch for each vertex and then parameterizeit into a 2D grid (or image). We then embed identified shape descriptorsinto the 2D grids which are further fed into the powerful 2D Convolu-tional Neural Network for regressing corresponding semantic labels (e.g.,head, torso). Experiments demonstrate that our method is effective inhuman segmentation, and achieves state-of-the-art accuracy.

Takumi Nakane, Xuequan Lu, Chao Zhang

In evolutionary algorithms, genetic operators iteratively generate new offspring which constitute a potentially valuable set of search history. To boost the performance of crossover in real-coded genetic algorithm (RCGA), in this paper we propose to exploit the search history cached so far in an online style during the iteration. Specifically, survivor individuals over past few generations are collected and stored in the archive to form the search history. We introduce a simple yet effective crossover model driven by the search history (abbreviated as SHX). In particular, the search history is clustered and each cluster is assigned a score for SHX. In essence, the proposed SHX is a data-driven method which exploits the search history to perform offspring selection after the offspring generation. Since no additional fitness evaluations are needed, SHX is favorable for the tasks with limited budget or expensive fitness evaluations. We experimentally verify the effectiveness of SHX over 4 benchmark functions. Quantitative results show that our SHX can significantly enhance the performance of RCGA, in terms of accuracy.

Wanzheng Zhu, Hongyu Gong, Jiaming Shen et al.

Set expansion aims to expand a small set of seed entities into a complete set of relevant entities. Most existing approaches assume the input seed set is unambiguous and completely ignore the multi-faceted semantics of seed entities. As a result, given the seed set {"Canon", "Sony", "Nikon"}, previous models return one mixed set of entities that are either Camera Brands or Japanese Companies. In this paper, we study the task of multi-faceted set expansion, which aims to capture all semantic facets in the seed set and return multiple sets of entities, one for each semantic facet. We propose an unsupervised framework, FUSE, which consists of three major components: (1) facet discovery module: identifies all semantic facets of each seed entity by extracting and clustering its skip-grams, and (2) facet fusion module: discovers shared semantic facets of the entire seed set by an optimization formulation, and (3) entity expansion module: expands each semantic facet by utilizing a masked language model with pre-trained BERT models. Extensive experiments demonstrate that FUSE can accurately identify multiple semantic facets of the seed set and generate quality entities for each facet.

Lang Huang, Yuhui Yuan, Jianyuan Guo et al.

In this paper, we present a so-called interlaced sparse self-attention approach to improve the efficiency of the \emph{self-attention} mechanism for semantic segmentation. The main idea is that we factorize the dense affinity matrix as the product of two sparse affinity matrices. There are two successive attention modules each estimating a sparse affinity matrix. The first attention module is used to estimate the affinities within a subset of positions that have long spatial interval distances and the second attention module is used to estimate the affinities within a subset of positions that have short spatial interval distances. These two attention modules are designed so that each position is able to receive the information from all the other positions. In contrast to the original self-attention module, our approach decreases the computation and memory complexity substantially especially when processing high-resolution feature maps. We empirically verify the effectiveness of our approach on six challenging semantic segmentation benchmarks.

Yi Zhang, Chao Zhang, Takuya Akashi

We propose a novel multi-scale template matching method which is robust against both scaling and rotation in unconstrained environments. The key component behind is a similarity measure referred to as scalable diversity similarity (SDS). Specifically, SDS exploits bidirectional diversity of the nearest neighbor (NN) matches between two sets of points. To address the scale-robustness of the similarity measure, local appearance and rank information are jointly used for the NN search. Furthermore, by introducing penalty term on the scale change, and polar radius term into the similarity measure, SDS is shown to be a well-performing similarity measure against overall size and rotation changes, as well as non-rigid geometric deformations, background clutter, and occlusions. The properties of SDS are statistically justified, and experiments on both synthetic and real-world data show that SDS can significantly outperform state-of-the-art methods.

Takumi Nakane, Takuya Akashi, Xuequan Lu et al.

We propose a novel genetic algorithm to solve the image deformation estimation problem by preserving the genetic diversity. As a classical problem, there is always a trade-off between the complexity of deformation models and the difficulty of parameters search in image deformation. 2D cubic B-spline surface is a highly free-form deformation model and is able to handle complex deformations such as fluid image distortions. However, it is challenging to estimate an apposite global solution. To tackle this problem, we develop a genetic operation named probabilistic bitwise operation (PBO) to replace the crossover and mutation operations, which can preserve the diversity during generation iteration and achieve better coverage ratio of the solution space. Furthermore, a selection strategy named annealing selection is proposed to control the convergence. Qualitative and quantitative results on synthetic data show the effectiveness of our method.

Bold Naranchimeg, Chao Zhang, Takuya Akashi

In recent decade, many state-of-the-art algorithms on image classification as well as audio classification have achieved noticeable successes with the development of deep convolutional neural network (CNN). However, most of the works only exploit single type of training data. In this paper, we present a study on classifying bird species by exploiting the combination of both visual (images) and audio (sounds) data using CNN, which has been sparsely treated so far. Specifically, we propose CNN-based multimodal learning models in three types of fusion strategies (early, middle, late) to settle the issues of combining training data cross domains. The advantage of our proposed method lies on the fact that We can utilize CNN not only to extract features from image and audio data (spectrogram) but also to combine the features across modalities. In the experiment, we train and evaluate the network structure on a comprehensive CUB-200-2011 standard data set combing our originally collected audio data set with respect to the data species. We observe that a model which utilizes the combination of both data outperforms models trained with only an either type of data. We also show that transfer learning can significantly increase the classification performance.

Yiwen Guo, Chao Zhang, Changshui Zhang et al.

Deep neural networks (DNNs) are computationally/memory-intensive and vulnerable to adversarial attacks, making them prohibitive in some real-world applications. By converting dense models into sparse ones, pruning appears to be a promising solution to reducing the computation/memory cost. This paper studies classification models, especially DNN-based ones, to demonstrate that there exists intrinsic relationships between their sparsity and adversarial robustness. Our analyses reveal, both theoretically and empirically, that nonlinear DNN-based classifiers behave differently under $l_2$ attacks from some linear ones. We further demonstrate that an appropriately higher model sparsity implies better robustness of nonlinear DNNs, whereas over-sparsified models can be more difficult to resist adversarial examples.

Felipe Petroski Such, Shagan Sah, Miguel Dominguez et al.

Convolutional Neural Networks (CNNs) have recently led to incredible breakthroughs on a variety of pattern recognition problems. Banks of finite impulse response filters are learned on a hierarchy of layers, each contributing more abstract information than the previous layer. The simplicity and elegance of the convolutional filtering process makes them perfect for structured problems such as image, video, or voice, where vertices are homogeneous in the sense of number, location, and strength of neighbors. The vast majority of classification problems, for example in the pharmaceutical, homeland security, and financial domains are unstructured. As these problems are formulated into unstructured graphs, the heterogeneity of these problems, such as number of vertices, number of connections per vertex, and edge strength, cannot be tackled with standard convolutional techniques. We propose a novel neural learning framework that is capable of handling both homogeneous and heterogeneous data, while retaining the benefits of traditional CNN successes. Recently, researchers have proposed variations of CNNs that can handle graph data. In an effort to create learnable filter banks of graphs, these methods either induce constraints on the data or require preprocessing. As opposed to spectral methods, our framework, which we term Graph-CNNs, defines filters as polynomials of functions of the graph adjacency matrix. Graph-CNNs can handle both heterogeneous and homogeneous graph data, including graphs having entirely different vertex or edge sets. We perform experiments to validate the applicability of Graph-CNNs to a variety of structured and unstructured classification problems and demonstrate state-of-the-art results on document and molecule classification problems.