arXiv subjects

Toshihiko Yamasaki

Publications and source records attributed to Toshihiko Yamasaki.

At least 73 records · Page 4Linked to original sources

arXiv2023

MetaMixer: A Regularization Strategy for Online Knowledge Distillation

Online knowledge distillation (KD) has received increasing attention in recent years. However, while most existing online KD methods focus on developing complicated model structures and training strategies to improve the distillation of high-level knowledge like probability distribution, the effects of the multi-level knowledge in the online KD are greatly overlooked, especially the low-level knowledge. Thus, to provide a novel viewpoint to online KD, we propose MetaMixer, a regularization strategy that can strengthen the distillation by combining the low-level knowledge that impacts the localization capability of the networks, and high-level knowledge that focuses on the whole image. Experiments under different conditions show that MetaMixer can achieve significant performance gains over state-of-the-art methods.

cs.CV↗

arXiv2023

Toward Extremely Lightweight Distracted Driver Recognition With Distillation-Based Neural Architecture Search and Knowledge Transfer

The number of traffic accidents has been continuously increasing in recent years worldwide. Many accidents are caused by distracted drivers, who take their attention away from driving. Motivated by the success of Convolutional Neural Networks (CNNs) in computer vision, many researchers developed CNN-based algorithms to recognize distracted driving from a dashcam and warn the driver against unsafe behaviors. However, current models have too many parameters, which is unfeasible for vehicle-mounted computing. This work proposes a novel knowledge-distillation-based framework to solve this problem. The proposed framework first constructs a high-performance teacher network by progressively strengthening the robustness to illumination changes from shallow to deep layers of a CNN. Then, the teacher network is used to guide the architecture searching process of a student network through knowledge distillation. After that, we use the teacher network again to transfer knowledge to the student network by knowledge distillation. Experimental results on the Statefarm Distracted Driver Detection Dataset and AUC Distracted Driver Dataset show that the proposed approach is highly effective for recognizing distracted driving behaviors from photos: (1) the teacher network's accuracy surpasses the previous best accuracy; (2) the student network achieves very high accuracy with only 0.42M parameters (around 55% of the previous most lightweight model). Furthermore, the student network architecture can be extended to a spatial-temporal 3D CNN for recognizing distracted driving from video clips. The 3D student network largely surpasses the previous best accuracy with only 2.03M parameters on the Drive&Act Dataset. The source code is available at https://github.com/Dichao-Liu/Lightweight_Distracted_Driver_Recognition_with_Distillation-Based_NAS_and_Knowledge_Transfer.

cs.CV↗

arXiv2022

Semi-supervised Fashion Compatibility Prediction by Color Distortion Prediction

Supervised learning methods have been suffering from the fact that a large-scale labeled dataset is mandatory, which is difficult to obtain. This has been a more significant issue for fashion compatibility prediction because compatibility aims to capture people's perception of aesthetics, which are sparse and changing. Thus, the labeled dataset may become outdated quickly due to fast fashion. Moreover, labeling the dataset always needs some expert knowledge; at least they should have a good sense of aesthetics. However, there are limited self/semi-supervised learning techniques in this field. In this paper, we propose a general color distortion prediction task forcing the baseline to recognize low-level image information to learn more discriminative representation for fashion compatibility prediction. Specifically, we first propose to distort the image by adjusting the image color balance, contrast, sharpness, and brightness. Then, we propose adding Gaussian noise to the distorted image before passing them to the convolutional neural network (CNN) backbone to learn a probability distribution over all possible distortions. The proposed pretext task is adopted in the state-of-the-art methods in fashion compatibility and shows its effectiveness in improving these methods' ability in extracting better feature representations. Applying the proposed pretext task to the baseline can consistently outperform the original baseline.

cs.CV↗

arXiv2022

Green Hierarchical Vision Transformer for Masked Image Modeling

We present an efficient approach for Masked Image Modeling (MIM) with hierarchical Vision Transformers (ViTs), allowing the hierarchical ViTs to discard masked patches and operate only on the visible ones. Our approach consists of three key designs. First, for window attention, we propose a Group Window Attention scheme following the Divide-and-Conquer strategy. To mitigate the quadratic complexity of the self-attention w.r.t. the number of patches, group attention encourages a uniform partition that visible patches within each local window of arbitrary size can be grouped with equal size, where masked self-attention is then performed within each group. Second, we further improve the grouping strategy via the Dynamic Programming algorithm to minimize the overall computation cost of the attention on the grouped patches. Third, as for the convolution layers, we convert them to the Sparse Convolution that works seamlessly with the sparse data, i.e., the visible patches in MIM. As a result, MIM can now work on most, if not all, hierarchical ViTs in a green and efficient way. For example, we can train the hierarchical ViTs, e.g., Swin Transformer and Twins Transformer, about 2.7$\times$ faster and reduce the GPU memory usage by 70%, while still enjoying competitive performance on ImageNet classification and the superiority on downstream COCO object detection benchmarks. Code and pre-trained models have been made publicly available at https://github.com/LayneH/GreenMIM.

cs.CV↗

arXiv2022

Fine-Grained Image Style Transfer with Visual Transformers

With the development of the convolutional neural network, image style transfer has drawn increasing attention. However, most existing approaches adopt a global feature transformation to transfer style patterns into content images (e.g., AdaIN and WCT). Such a design usually destroys the spatial information of the input images and fails to transfer fine-grained style patterns into style transfer results. To solve this problem, we propose a novel STyle TRansformer (STTR) network which breaks both content and style images into visual tokens to achieve a fine-grained style transformation. Specifically, two attention mechanisms are adopted in our STTR. We first propose to use self-attention to encode content and style tokens such that similar tokens can be grouped and learned together. We then adopt cross-attention between content and style tokens that encourages fine-grained style transformations. To compare STTR with existing approaches, we conduct user studies on Amazon Mechanical Turk (AMT), which are carried out with 50 human subjects with 1,000 votes in total. Extensive evaluations demonstrate the effectiveness and efficiency of the proposed STTR in generating visually pleasing style transfer results.

cs.CV↗

arXiv2022

Improving Robustness to Out-of-Distribution Data by Frequency-based Augmentation

Although Convolutional Neural Networks (CNNs) have high accuracy in image recognition, they are vulnerable to adversarial examples and out-of-distribution data, and the difference from human recognition has been pointed out. In order to improve the robustness against out-of-distribution data, we present a frequency-based data augmentation technique that replaces the frequency components with other images of the same class. When the training data are CIFAR10 and the out-of-distribution data are SVHN, the Area Under Receiver Operating Characteristic (AUROC) curve of the model trained with the proposed method increases from 89.22\% to 98.15\%, and further increased to 98.59\% when combined with another data augmentation method. Furthermore, we experimentally demonstrate that the robust model for out-of-distribution data uses a lot of high-frequency components of the image.

cs.CV↗

arXiv2022

Prediction of Seismic Intensity Distributions Using Neural Networks

The ground motion prediction equation is commonly used to predict the seismic intensity distribution. However, it is not easy to apply this method to seismic distributions affected by underground plate structures, which are commonly known as abnormal seismic distributions. This study proposes a hybrid of regression and classification approaches using neural networks. The proposed model treats the distributions as 2-dimensional data like an image. Our method can accurately predict seismic intensity distributions, even abnormal distributions.

cs.CV↗

arXiv2022

Superclass Adversarial Attack

Adversarial attacks have only focused on changing the predictions of the classifier, but their danger greatly depends on how the class is mistaken. For example, when an automatic driving system mistakes a Persian cat for a Siamese cat, it is hardly a problem. However, if it mistakes a cat for a 120km/h minimum speed sign, serious problems can arise. As a stepping stone to more threatening adversarial attacks, we consider the superclass adversarial attack, which causes misclassification of not only fine classes, but also superclasses. We conducted the first comprehensive analysis of superclass adversarial attacks (an existing and 19 new methods) in terms of accuracy, speed, and stability, and identified several strategies to achieve better performance. Although this study is aimed at superclass misclassification, the findings can be applied to other problem settings involving multiple classes, such as top-k and multi-label classification attacks.

cs.CV↗

arXiv2022

SAT: Self-adaptive training for fashion compatibility prediction

This paper presents a self-adaptive training (SAT) model for fashion compatibility prediction. It focuses on the learning of some hard items, such as those that share similar color, texture, and pattern features but are considered incompatible due to the aesthetics or temporal shifts. Specifically, we first design a method to define hard outfits and a difficulty score (DS) is defined and assigned to each outfit based on the difficulty in recommending an item for it. Then, we propose a self-adaptive triplet loss (SATL), where the DS of the outfit is considered. Finally, we propose a very simple conditional similarity network combining the proposed SATL to achieve the learning of hard items in the fashion compatibility prediction. Experiments on the publicly available Polyvore Outfits and Polyvore Outfits-D datasets demonstrate our SAT's effectiveness in fashion compatibility prediction. Besides, our SATL can be easily extended to other conditional similarity networks to improve their performance.

cs.CV↗

arXiv2022

Detecting Deepfakes with Self-Blended Images

In this paper, we present novel synthetic training data called self-blended images (SBIs) to detect deepfakes. SBIs are generated by blending pseudo source and target images from single pristine images, reproducing common forgery artifacts (e.g., blending boundaries and statistical inconsistencies between source and target images). The key idea behind SBIs is that more general and hardly recognizable fake samples encourage classifiers to learn generic and robust representations without overfitting to manipulation-specific artifacts. We compare our approach with state-of-the-art methods on FF++, CDF, DFD, DFDC, DFDCP, and FFIW datasets by following the standard cross-dataset and cross-manipulation protocols. Extensive experiments show that our method improves the model generalization to unknown manipulations and scenes. In particular, on DFDC and DFDCP where existing methods suffer from the domain gap between the training and test sets, our approach outperforms the baseline by 4.90% and 11.78% points in the cross-dataset evaluation, respectively.

cs.CV↗

arXiv2022

Learning Where to Learn in Cross-View Self-Supervised Learning

Self-supervised learning (SSL) has made enormous progress and largely narrowed the gap with the supervised ones, where the representation learning is mainly guided by a projection into an embedding space. During the projection, current methods simply adopt uniform aggregation of pixels for embedding; however, this risks involving object-irrelevant nuisances and spatial misalignment for different augmentations. In this paper, we present a new approach, Learning Where to Learn (LEWEL), to adaptively aggregate spatial information of features, so that the projected embeddings could be exactly aligned and thus guide the feature learning better. Concretely, we reinterpret the projection head in SSL as a per-pixel projection and predict a set of spatial alignment maps from the original features by this weight-sharing projection head. A spectrum of aligned embeddings is thus obtained by aggregating the features with spatial weighting according to these alignment maps. As a result of this adaptive alignment, we observe substantial improvements on both image-level prediction and dense prediction at the same time: LEWEL improves MoCov2 by 1.6%/1.3%/0.5%/0.4% points, improves BYOL by 1.3%/1.3%/0.7%/0.6% points, on ImageNet linear/semi-supervised classification, Pascal VOC semantic segmentation, and object detection, respectively.

cs.CV↗

arXiv2022

Are DNNs fooled by extremely unrecognizable images?

Fooling images are a potential threat to deep neural networks (DNNs). These images are not recognizable to humans as natural objects, such as dogs and cats, but are misclassified by DNNs as natural-object classes with high confidence scores. Despite their original design concept, existing fooling images retain some features that are characteristic of the target objects if looked into closely. Hence, DNNs can react to these features. In this paper, we address the question of whether there can be fooling images with no characteristic pattern of natural objects locally or globally. As a minimal case, we introduce single-color images with a few pixels altered, called sparse fooling images (SFIs). We first prove that SFIs always exist under mild conditions for linear and nonlinear models and reveal that complex models are more likely to be vulnerable to SFI attacks. With two SFI generation methods, we demonstrate that in deeper layers, SFIs end up with similar features to those of natural images, and consequently, fool DNNs successfully. Among other layers, we discovered that the max pooling layer causes the vulnerability against SFIs. The defense against SFIs and transferability are also discussed. This study highlights the new vulnerability of DNNs by introducing a novel class of images that distributes extremely far from natural images.

cs.CV↗

arXiv2022

Subjective Functionality and Comfort Prediction for Apartment Floor Plans and Its Application to Intuitive Searches

This study presents a new user experience in apartment searches using functionality and comfort as query items. This study has three technical contributions. First, we present a new dataset on the perceived functionality and comfort scores of residential floor plans using nine question statements about the level of comfort, openness, privacy, etc. Second, we propose an algorithm to predict the scores from the floor plan images. Lastly, we implement a new apartment search system and conduct a large-scale usability study using crowdsourcing. The experimental results show that our apartment search system can provide a better user experience. To the best of our knowledge, this study is the first work to propose a highly accurate prediction model for the subjective functionality and comfort of apartments using machine learning.

cs.HC↗

arXiv2021

Edge-Level Explanations for Graph Neural Networks by Extending Explainability Methods for Convolutional Neural Networks

Graph Neural Networks (GNNs) are deep learning models that take graph data as inputs, and they are applied to various tasks such as traffic prediction and molecular property prediction. However, owing to the complexity of the GNNs, it has been difficult to analyze which parts of inputs affect the GNN model's outputs. In this study, we extend explainability methods for Convolutional Neural Networks (CNNs), such as Local Interpretable Model-Agnostic Explanations (LIME), Gradient-Based Saliency Maps, and Gradient-Weighted Class Activation Mapping (Grad-CAM) to GNNs, and predict which edges in the input graphs are important for GNN decisions. The experimental results indicate that the LIME-based approach is the most efficient explainability method for multiple tasks in the real-world situation, outperforming even the state-of-the-art method in GNN explainability.

cs.LG↗

arXiv2021

Location Prediction via Bi-direction Speculation and Dual-level Association

Location prediction is of great importance in location-based applications for the construction of the smart city. To our knowledge, existing models for location prediction focus on users' preferences on POIs from the perspective of the human side. However, modeling users' interests from the historical trajectory is still limited by the data sparsity. Additionally, most of existing methods predict the next location according to the individual data independently, but the data sparsity makes it difficult to mine explicit mobility patterns or capture the casual behavior for each user. To address the issues above, we propose a novel Bi-direction Speculation and Dual-level Association method (BSDA), which considers both users' interests in POIs and POIs' appeal to users. Furthermore, we develop the cross-user and cross-POI association to alleviate the data sparsity by similar users and POIs to enrich the candidates. Experimental results on two public datasets demonstrate that BSDA achieves significant improvements over state-of-the-art methods.

cs.SI↗

arXiv2021

Unsupervised Video Person Re-identification via Noise and Hard frame Aware Clustering

Unsupervised video-based person re-identification (re-ID) methods extract richer features from video tracklets than image-based ones. The state-of-the-art methods utilize clustering to obtain pseudo-labels and train the models iteratively. However, they underestimate the influence of two kinds of frames in the tracklet: 1) noise frames caused by detection errors or heavy occlusions exist in the tracklet, which may be allocated with unreliable labels during clustering; 2) the tracklet also contains hard frames caused by pose changes or partial occlusions, which are difficult to distinguish but informative. This paper proposes a Noise and Hard frame Aware Clustering (NHAC) method. NHAC consists of a graph trimming module and a node re-sampling module. The graph trimming module obtains stable graphs by removing noise frame nodes to improve the clustering accuracy. The node re-sampling module enhances the training of hard frame nodes to learn rich tracklet information. Experiments conducted on two video-based datasets demonstrate the effectiveness of the proposed NHAC under the unsupervised re-ID setting.

cs.CV↗

arXiv2021

Pretext-Contrastive Learning: Toward Good Practices in Self-supervised Video Representation Leaning

Recently, pretext-task based methods are proposed one after another in self-supervised video feature learning. Meanwhile, contrastive learning methods also yield good performance. Usually, new methods can beat previous ones as claimed that they could capture "better" temporal information. However, there exist setting differences among them and it is hard to conclude which is better. It would be much more convincing in comparison if these methods have reached as closer to their performance limits as possible. In this paper, we start from one pretext-task baseline, exploring how far it can go by combining it with contrastive learning, data pre-processing, and data augmentation. A proper setting has been found from extensive experiments, with which huge improvements over the baselines can be achieved, indicating a joint optimization framework can boost both pretext task and contrastive learning. We denote the joint optimization framework as Pretext-Contrastive Learning (PCL). The other two pretext task baselines are used to validate the effectiveness of PCL. And we can easily outperform current state-of-the-art methods in the same training manner, showing the effectiveness and the generality of our proposal. It is convenient to treat PCL as a standard training strategy and apply it to many other works in self-supervised video feature learning.

cs.CV↗

arXiv2021

Learning from Synthetic Shadows for Shadow Detection and Removal

Shadow removal is an essential task in computer vision and computer graphics. Recent shadow removal approaches all train convolutional neural networks (CNN) on real paired shadow/shadow-free or shadow/shadow-free/mask image datasets. However, obtaining a large-scale, diverse, and accurate dataset has been a big challenge, and it limits the performance of the learned models on shadow images with unseen shapes/intensities. To overcome this challenge, we present SynShadow, a novel large-scale synthetic shadow/shadow-free/matte image triplets dataset and a pipeline to synthesize it. We extend a physically-grounded shadow illumination model and synthesize a shadow image given an arbitrary combination of a shadow-free image, a matte image, and shadow attenuation parameters. Owing to the diversity, quantity, and quality of SynShadow, we demonstrate that shadow removal models trained on SynShadow perform well in removing shadows with diverse shapes and intensities on some challenging benchmarks. Furthermore, we show that merely fine-tuning from a SynShadow-pre-trained model improves existing shadow detection and removal models. Codes are publicly available at https://github.com/naoto0804/SynShadow.

cs.CV↗