Pub Date : 2021-06-01DOI: 10.1109/CVPR46437.2021.00822
Jiayan Qiu, Yiding Yang, Xinchao Wang, D. Tao
What scene elements, if any, are indispensable for recognizing a scene? We strive to answer this question through the lens of an exotic learning scheme. Our goal is to identify a collection of such pivotal elements, which we term as Scene Essence, to be those that would alter scene recognition if taken out from the scene. To this end, we devise a novel approach that learns to partition the scene objects into two groups, essential ones and minor ones, under the supervision that if only the essential ones are kept while the minor ones are erased in the input image, a scene recognizer would preserve its original prediction. Specifically, we introduce a learnable graph neural network (GNN) for labelling scene objects, based on which the minor ones are wiped off by an off-the-shelf image inpainter. The features of the inpainted image derived in this way, together with those learned from the GNN with the minor-object nodes pruned, are expected to fool the scene discriminator. Both subjective and objective evaluations on Places365, SUN397, and MIT67 datasets demonstrate that, the learned Scene Essence yields a visually plausible image that convincingly retains the original scene category.
{"title":"Scene Essence","authors":"Jiayan Qiu, Yiding Yang, Xinchao Wang, D. Tao","doi":"10.1109/CVPR46437.2021.00822","DOIUrl":"https://doi.org/10.1109/CVPR46437.2021.00822","url":null,"abstract":"What scene elements, if any, are indispensable for recognizing a scene? We strive to answer this question through the lens of an exotic learning scheme. Our goal is to identify a collection of such pivotal elements, which we term as Scene Essence, to be those that would alter scene recognition if taken out from the scene. To this end, we devise a novel approach that learns to partition the scene objects into two groups, essential ones and minor ones, under the supervision that if only the essential ones are kept while the minor ones are erased in the input image, a scene recognizer would preserve its original prediction. Specifically, we introduce a learnable graph neural network (GNN) for labelling scene objects, based on which the minor ones are wiped off by an off-the-shelf image inpainter. The features of the inpainted image derived in this way, together with those learned from the GNN with the minor-object nodes pruned, are expected to fool the scene discriminator. Both subjective and objective evaluations on Places365, SUN397, and MIT67 datasets demonstrate that, the learned Scene Essence yields a visually plausible image that convincingly retains the original scene category.","PeriodicalId":339646,"journal":{"name":"2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)","volume":"122 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127982689","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2021-06-01DOI: 10.1109/CVPR46437.2021.01048
Yao Lee, Kuan-Wei Tseng, Yu-Ta Chen, Chien-Cheng Chen, Chu-Song Chen, Y. Hung
Video stabilization is an essential component of visual quality enhancement. Early methods rely on feature tracking to recover either 2D or 3D frame motion, which suffer from the robustness of local feature extraction and tracking in shaky videos. Recently, learning-based methods seek to find frame transformations with high-level information via deep neural networks to overcome the robustness issue of feature tracking. Nevertheless, to our best knowledge, no learning-based methods leverage 3D cues for the transformation inference yet; hence they would lead to artifacts on complex scene-depth scenarios. In this paper, we propose Deep3D Stabilizer, a novel 3D depth-based learning method for video stabilization. We take advantage of the recent self-supervised framework on jointly learning depth and camera ego-motion estimation on raw videos. Our approach requires no data for pre-training but stabilizes the input video via 3D reconstruction directly. The rectification stage incorporates the 3D scene depth and camera motion to smooth the camera trajectory and synthesize the stabilized video. Unlike most one-size-fits-all learning-based methods, our smoothing algorithm allows users to manipulate the stability of a video efficiently. Experimental results on challenging benchmarks show that the proposed solution consistently outperforms the state-of-the-art methods on almost all motion categories.
{"title":"3D Video Stabilization with Depth Estimation by CNN-based Optimization","authors":"Yao Lee, Kuan-Wei Tseng, Yu-Ta Chen, Chien-Cheng Chen, Chu-Song Chen, Y. Hung","doi":"10.1109/CVPR46437.2021.01048","DOIUrl":"https://doi.org/10.1109/CVPR46437.2021.01048","url":null,"abstract":"Video stabilization is an essential component of visual quality enhancement. Early methods rely on feature tracking to recover either 2D or 3D frame motion, which suffer from the robustness of local feature extraction and tracking in shaky videos. Recently, learning-based methods seek to find frame transformations with high-level information via deep neural networks to overcome the robustness issue of feature tracking. Nevertheless, to our best knowledge, no learning-based methods leverage 3D cues for the transformation inference yet; hence they would lead to artifacts on complex scene-depth scenarios. In this paper, we propose Deep3D Stabilizer, a novel 3D depth-based learning method for video stabilization. We take advantage of the recent self-supervised framework on jointly learning depth and camera ego-motion estimation on raw videos. Our approach requires no data for pre-training but stabilizes the input video via 3D reconstruction directly. The rectification stage incorporates the 3D scene depth and camera motion to smooth the camera trajectory and synthesize the stabilized video. Unlike most one-size-fits-all learning-based methods, our smoothing algorithm allows users to manipulate the stability of a video efficiently. Experimental results on challenging benchmarks show that the proposed solution consistently outperforms the state-of-the-art methods on almost all motion categories.","PeriodicalId":339646,"journal":{"name":"2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)","volume":"29 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115918542","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2021-06-01DOI: 10.1109/CVPR46437.2021.01461
A. Chadha, Y. Andreopoulos
We introduce the concept of rate-aware deep perceptual preprocessing (DPP) for video encoding. DPP makes a single pass over each input frame in order to enhance its visual quality when the video is to be compressed with any codec at any bitrate. The resulting bitstreams can be decoded and displayed at the client side without any post-processing component. DPP comprises a convolutional neural network that is trained via a composite set of loss functions that incorporates: (i) a perceptual loss based on a trained no-reference image quality assessment model, (ii) a reference-based fidelity loss expressing L1 and structural similarity aspects, (iii) a motion-based rate loss via block-based transform, quantization and entropy estimates that converts the essential components of standard hybrid video encoder designs into a trainable framework. Extensive testing using multiple quality metrics and AVC, AV1 and VVC encoders shows that DPP+encoder reduces, on average, the bitrate of the corresponding encoder by 11%. This marks the first time a server-side neural processing component achieves such savings over the state-of-the-art in video coding.
{"title":"Deep Perceptual Preprocessing for Video Coding","authors":"A. Chadha, Y. Andreopoulos","doi":"10.1109/CVPR46437.2021.01461","DOIUrl":"https://doi.org/10.1109/CVPR46437.2021.01461","url":null,"abstract":"We introduce the concept of rate-aware deep perceptual preprocessing (DPP) for video encoding. DPP makes a single pass over each input frame in order to enhance its visual quality when the video is to be compressed with any codec at any bitrate. The resulting bitstreams can be decoded and displayed at the client side without any post-processing component. DPP comprises a convolutional neural network that is trained via a composite set of loss functions that incorporates: (i) a perceptual loss based on a trained no-reference image quality assessment model, (ii) a reference-based fidelity loss expressing L1 and structural similarity aspects, (iii) a motion-based rate loss via block-based transform, quantization and entropy estimates that converts the essential components of standard hybrid video encoder designs into a trainable framework. Extensive testing using multiple quality metrics and AVC, AV1 and VVC encoders shows that DPP+encoder reduces, on average, the bitrate of the corresponding encoder by 11%. This marks the first time a server-side neural processing component achieves such savings over the state-of-the-art in video coding.","PeriodicalId":339646,"journal":{"name":"2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)","volume":"149 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132225825","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2021-06-01DOI: 10.1109/CVPR46437.2021.00483
H. Kim, Sunghun Joung, Ig-Jae Kim, K. Sohn
Existing person search methods integrate person detection and re-identification (re-ID) module into a unified system. Though promising results have been achieved, the misalignment problem, which commonly occurs in person search, limits the discriminative feature representation for re-ID. To overcome this limitation, we introduce a novel framework to learn the discriminative representation by utilizing prototype in OIM loss. Unlike conventional methods using prototype as a representation of person identity, we utilize it as guidance to allow the attention network to consistently highlight multiple instances across different poses. Moreover, we propose a new prototype update scheme with adaptive momentum to increase the discriminative ability across different instances. Extensive ablation experiments demonstrate that our method can significantly enhance the feature discriminative power, outperforming the state-of-the-art results on two person search benchmarks including CUHK-SYSU and PRW.
{"title":"Prototype-Guided Saliency Feature Learning for Person Search","authors":"H. Kim, Sunghun Joung, Ig-Jae Kim, K. Sohn","doi":"10.1109/CVPR46437.2021.00483","DOIUrl":"https://doi.org/10.1109/CVPR46437.2021.00483","url":null,"abstract":"Existing person search methods integrate person detection and re-identification (re-ID) module into a unified system. Though promising results have been achieved, the misalignment problem, which commonly occurs in person search, limits the discriminative feature representation for re-ID. To overcome this limitation, we introduce a novel framework to learn the discriminative representation by utilizing prototype in OIM loss. Unlike conventional methods using prototype as a representation of person identity, we utilize it as guidance to allow the attention network to consistently highlight multiple instances across different poses. Moreover, we propose a new prototype update scheme with adaptive momentum to increase the discriminative ability across different instances. Extensive ablation experiments demonstrate that our method can significantly enhance the feature discriminative power, outperforming the state-of-the-art results on two person search benchmarks including CUHK-SYSU and PRW.","PeriodicalId":339646,"journal":{"name":"2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130431733","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2021-06-01DOI: 10.1109/CVPR46437.2021.01311
Pengyu Li, Biao Wang, Lei Zhang
Recently, deep face recognition has achieved significant progress because of Convolutional Neural Networks (CNNs) and large-scale datasets. However, training CNNs on a large-scale face recognition dataset with limited computational resources is still a challenge. This is because the classification paradigm needs to train a fully-connected layer as the category classifier, and its parameters will be in the hundreds of millions if the training dataset contains millions of identities. This requires many computational resources, such as GPU memory. The metric learning paradigm is an economical computation method, but its performance is greatly inferior to that of the classification paradigm. To address this challenge, we propose a simple but effective CNN layer called the Virtual fully-connected (Virtual FC) layer to reduce the computational consumption of the classification paradigm. Without bells and whistles, the proposed Virtual FC reduces the parameters by more than 100 times with respect to the fully-connected layer and achieves competitive performance on mainstream face recognition evaluation datasets. Moreover, the performance of our Virtual FC layer on the evaluation datasets is superior to that of the metric learning paradigm by a significant margin. Our code will be released in hopes of disseminating our idea to other domains1.
{"title":"Virtual Fully-Connected Layer: Training a Large-Scale Face Recognition Dataset with Limited Computational Resources","authors":"Pengyu Li, Biao Wang, Lei Zhang","doi":"10.1109/CVPR46437.2021.01311","DOIUrl":"https://doi.org/10.1109/CVPR46437.2021.01311","url":null,"abstract":"Recently, deep face recognition has achieved significant progress because of Convolutional Neural Networks (CNNs) and large-scale datasets. However, training CNNs on a large-scale face recognition dataset with limited computational resources is still a challenge. This is because the classification paradigm needs to train a fully-connected layer as the category classifier, and its parameters will be in the hundreds of millions if the training dataset contains millions of identities. This requires many computational resources, such as GPU memory. The metric learning paradigm is an economical computation method, but its performance is greatly inferior to that of the classification paradigm. To address this challenge, we propose a simple but effective CNN layer called the Virtual fully-connected (Virtual FC) layer to reduce the computational consumption of the classification paradigm. Without bells and whistles, the proposed Virtual FC reduces the parameters by more than 100 times with respect to the fully-connected layer and achieves competitive performance on mainstream face recognition evaluation datasets. Moreover, the performance of our Virtual FC layer on the evaluation datasets is superior to that of the metric learning paradigm by a significant margin. Our code will be released in hopes of disseminating our idea to other domains1.","PeriodicalId":339646,"journal":{"name":"2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)","volume":"26 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"134507730","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2021-06-01DOI: 10.1109/CVPR46437.2021.00144
Pei Wang, Kabir Nagrecha, N. Vasconcelos
The problem of machine teaching is considered. A new formulation is proposed under the assumption of an optimal student, where optimality is defined in the usual machine learning sense of empirical risk minimization. This is a sensible assumption for machine learning students and for human students in crowdsourcing platforms, who tend to perform at least as well as machine learning systems. It is shown that, if allowed unbounded effort, the optimal student always learns the optimal predictor for a classification task. Hence, the role of the optimal teacher is to select the teaching set that minimizes student effort. This is formulated as a problem of functional optimization where, at each teaching iteration, the teacher seeks to align the steepest descent directions of the risk of (1) the teaching set and (2) entire example population. The optimal teacher, denoted MaxGrad, is then shown to maximize the gradient of the risk on the set of new examples selected per iteration. MaxGrad teaching algorithms are finally provided for both binary and multiclass tasks, and shown to have some similarities with boosting algorithms. Experimental evaluations demonstrate the effectiveness of MaxGrad, which outperforms previous algorithms on the classification task, for both machine learning and human students from MTurk, by a substantial margin.
{"title":"Gradient-based Algorithms for Machine Teaching","authors":"Pei Wang, Kabir Nagrecha, N. Vasconcelos","doi":"10.1109/CVPR46437.2021.00144","DOIUrl":"https://doi.org/10.1109/CVPR46437.2021.00144","url":null,"abstract":"The problem of machine teaching is considered. A new formulation is proposed under the assumption of an optimal student, where optimality is defined in the usual machine learning sense of empirical risk minimization. This is a sensible assumption for machine learning students and for human students in crowdsourcing platforms, who tend to perform at least as well as machine learning systems. It is shown that, if allowed unbounded effort, the optimal student always learns the optimal predictor for a classification task. Hence, the role of the optimal teacher is to select the teaching set that minimizes student effort. This is formulated as a problem of functional optimization where, at each teaching iteration, the teacher seeks to align the steepest descent directions of the risk of (1) the teaching set and (2) entire example population. The optimal teacher, denoted MaxGrad, is then shown to maximize the gradient of the risk on the set of new examples selected per iteration. MaxGrad teaching algorithms are finally provided for both binary and multiclass tasks, and shown to have some similarities with boosting algorithms. Experimental evaluations demonstrate the effectiveness of MaxGrad, which outperforms previous algorithms on the classification task, for both machine learning and human students from MTurk, by a substantial margin.","PeriodicalId":339646,"journal":{"name":"2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)","volume":"219 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133959051","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2021-06-01DOI: 10.1109/CVPR46437.2021.00389
Marcus Valtonen Örnhag, J. Iglesias, Carl Olsson
Low rank inducing penalties have been proven to successfully uncover fundamental structures considered in computer vision and machine learning; however, such methods generally lead to non-convex optimization problems. Since the resulting objective is non-convex one often resorts to using standard splitting schemes such as Alternating Direction Methods of Multipliers (ADMM), or other subgradient methods, which exhibit slow convergence in the neighbourhood of a local minimum. We propose a method using second order methods, in particular the variable projection method (VarPro), by replacing the nonconvex penalties with a surrogate capable of converting the original objectives to differentiable equivalents. In this way we benefit from faster convergence.The bilinear framework is compatible with a large family of regularizers, and we demonstrate the benefits of our approach on real datasets for rigid and non-rigid structure from motion. The qualitative difference in reconstructions show that many popular non-convex objectives enjoy an advantage in transitioning to the proposed framework.1
{"title":"Bilinear Parameterization for Non-Separable Singular Value Penalties","authors":"Marcus Valtonen Örnhag, J. Iglesias, Carl Olsson","doi":"10.1109/CVPR46437.2021.00389","DOIUrl":"https://doi.org/10.1109/CVPR46437.2021.00389","url":null,"abstract":"Low rank inducing penalties have been proven to successfully uncover fundamental structures considered in computer vision and machine learning; however, such methods generally lead to non-convex optimization problems. Since the resulting objective is non-convex one often resorts to using standard splitting schemes such as Alternating Direction Methods of Multipliers (ADMM), or other subgradient methods, which exhibit slow convergence in the neighbourhood of a local minimum. We propose a method using second order methods, in particular the variable projection method (VarPro), by replacing the nonconvex penalties with a surrogate capable of converting the original objectives to differentiable equivalents. In this way we benefit from faster convergence.The bilinear framework is compatible with a large family of regularizers, and we demonstrate the benefits of our approach on real datasets for rigid and non-rigid structure from motion. The qualitative difference in reconstructions show that many popular non-convex objectives enjoy an advantage in transitioning to the proposed framework.1","PeriodicalId":339646,"journal":{"name":"2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)","volume":"34 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"134168173","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2021-06-01DOI: 10.1109/CVPR46437.2021.00623
N. Robidoux, L. E. G. Capel, Dongmin Seo, Avinash Sharma, Federico Ariza, Felix Heide
The real world is a 280 dB High Dynamic Range (HDR) world which imaging sensors cannot record in a single shot. HDR cameras acquire multiple measurements with different exposures, gains and photodiodes, from which an Image Signal Processor (ISP) reconstructs an HDR image. Dynamic scene HDR image recovery is an open challenge because of motion and because stitched captures have different noise characteristics, resulting in artifacts that ISPs must resolve in real time at double-digit megapixel resolutions. Traditionally, ISP settings used by downstream vision modules are chosen by domain experts; such frozen camera designs are then used for training data acquisition and supervised learning of downstream vision modules. We depart from this paradigm and formulate HDR ISP hyperparameter search as an end-to-end optimization problem, proposing a mixed 0th and 1st-order block coordinate descent optimizer that jointly learns sensor, ISP and detector network weights using RAW image data augmented with emulated SNR transition region artifacts. We assess the proposed method for human vision and image understanding. For automotive object detection, the method improves mAP and mAR by 33% over expert-tuning and 22% over state-of-the-art optimization methods, outperforming expert-tuned HDR imaging and vision pipelines in all HDR laboratory rig and field experiments.
{"title":"End-to-end High Dynamic Range Camera Pipeline Optimization","authors":"N. Robidoux, L. E. G. Capel, Dongmin Seo, Avinash Sharma, Federico Ariza, Felix Heide","doi":"10.1109/CVPR46437.2021.00623","DOIUrl":"https://doi.org/10.1109/CVPR46437.2021.00623","url":null,"abstract":"The real world is a 280 dB High Dynamic Range (HDR) world which imaging sensors cannot record in a single shot. HDR cameras acquire multiple measurements with different exposures, gains and photodiodes, from which an Image Signal Processor (ISP) reconstructs an HDR image. Dynamic scene HDR image recovery is an open challenge because of motion and because stitched captures have different noise characteristics, resulting in artifacts that ISPs must resolve in real time at double-digit megapixel resolutions. Traditionally, ISP settings used by downstream vision modules are chosen by domain experts; such frozen camera designs are then used for training data acquisition and supervised learning of downstream vision modules. We depart from this paradigm and formulate HDR ISP hyperparameter search as an end-to-end optimization problem, proposing a mixed 0th and 1st-order block coordinate descent optimizer that jointly learns sensor, ISP and detector network weights using RAW image data augmented with emulated SNR transition region artifacts. We assess the proposed method for human vision and image understanding. For automotive object detection, the method improves mAP and mAR by 33% over expert-tuning and 22% over state-of-the-art optimization methods, outperforming expert-tuned HDR imaging and vision pipelines in all HDR laboratory rig and field experiments.","PeriodicalId":339646,"journal":{"name":"2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)","volume":"126 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130910523","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2021-06-01DOI: 10.1109/CVPR46437.2021.00856
Shangqi Gao, Qi Han, Duo Li, Ming-Ming Cheng, Pai Peng
Batch Normalization (BatchNorm) has become the default component in modern neural networks to stabilize training. In BatchNorm, centering and scaling operations, along with mean and variance statistics, are utilized for feature standardization over the batch dimension. The batch dependency of BatchNorm enables stable training and better representation of the network, while inevitably ignores the representation differences among instances. We propose to add a simple yet effective feature calibration scheme into the centering and scaling operations of BatchNorm, enhancing the instance-specific representations with the negligible computational cost. The centering calibration strengthens informative features and reduces noisy features. The scaling calibration restricts the feature intensity to form a more stable feature distribution. Our proposed variant of BatchNorm, namely Representative BatchNorm, can be plugged into existing methods to boost the performance of various tasks such as classification, detection, and segmentation. The source code is available in http://mmcheng.net/rbn.
{"title":"Representative Batch Normalization with Feature Calibration","authors":"Shangqi Gao, Qi Han, Duo Li, Ming-Ming Cheng, Pai Peng","doi":"10.1109/CVPR46437.2021.00856","DOIUrl":"https://doi.org/10.1109/CVPR46437.2021.00856","url":null,"abstract":"Batch Normalization (BatchNorm) has become the default component in modern neural networks to stabilize training. In BatchNorm, centering and scaling operations, along with mean and variance statistics, are utilized for feature standardization over the batch dimension. The batch dependency of BatchNorm enables stable training and better representation of the network, while inevitably ignores the representation differences among instances. We propose to add a simple yet effective feature calibration scheme into the centering and scaling operations of BatchNorm, enhancing the instance-specific representations with the negligible computational cost. The centering calibration strengthens informative features and reduces noisy features. The scaling calibration restricts the feature intensity to form a more stable feature distribution. Our proposed variant of BatchNorm, namely Representative BatchNorm, can be plugged into existing methods to boost the performance of various tasks such as classification, detection, and segmentation. The source code is available in http://mmcheng.net/rbn.","PeriodicalId":339646,"journal":{"name":"2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)","volume":"80 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133096113","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2021-06-01DOI: 10.1109/CVPR46437.2021.01100
P. Ghosh, Nirat Saini, L. Davis, Abhinav Shrivastava
Fixed input graphs are a mainstay in approaches that utilize Graph Convolution Networks (GCNs) for knowledge transfer. The standard paradigm is to utilize relationships in the input graph to transfer information using GCNs from training to testing nodes in the graph; for example, the semi-supervised, zero-shot, and few-shot learning setups. We propose a generalized framework for learning and improving the input graph as part of the standard GCN-based learning setup. Moreover, we use additional constraints between similar and dissimilar neighbors for each node in the graph by applying triplet loss on the intermediate layer output. We present results of semi-supervised learning on Citeseer, Cora, and Pubmed benchmarking datasets, and zero/few-shot action recognition on UCF101 and HMDB51 datasets, significantly outperforming current approaches. We also present qualitative results visualizing the graph connections that our approach learns to update.
{"title":"Learning Graphs for Knowledge Transfer with Limited Labels","authors":"P. Ghosh, Nirat Saini, L. Davis, Abhinav Shrivastava","doi":"10.1109/CVPR46437.2021.01100","DOIUrl":"https://doi.org/10.1109/CVPR46437.2021.01100","url":null,"abstract":"Fixed input graphs are a mainstay in approaches that utilize Graph Convolution Networks (GCNs) for knowledge transfer. The standard paradigm is to utilize relationships in the input graph to transfer information using GCNs from training to testing nodes in the graph; for example, the semi-supervised, zero-shot, and few-shot learning setups. We propose a generalized framework for learning and improving the input graph as part of the standard GCN-based learning setup. Moreover, we use additional constraints between similar and dissimilar neighbors for each node in the graph by applying triplet loss on the intermediate layer output. We present results of semi-supervised learning on Citeseer, Cora, and Pubmed benchmarking datasets, and zero/few-shot action recognition on UCF101 and HMDB51 datasets, significantly outperforming current approaches. We also present qualitative results visualizing the graph connections that our approach learns to update.","PeriodicalId":339646,"journal":{"name":"2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)","volume":"22 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133494385","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：