Abstract
Safe motion planning in robotics requires planning into space which has been verified to be free of obstacles. However, obtaining such environment representations using lidars is challenging by virtue of the sparsity of their depth measurements. We present a learning-aided 3D lidar reconstruction framework that upsamples sparse lidar depth measurements with the aid of overlapping camera images so as to generate denser reconstructions with more definitively free space than can be achieved with the raw lidar measurements alone. We use a neural network with an encoder-decoder structure to predict dense depth images along with depth uncertainty estimates which are fused using a volumetric mapping system. We conduct experiments on real-world outdoor datasets captured using a handheld sensing device and a legged robot. Using input data from a 16-beam lidar mapping a building network, our experiments showed that the amount of estimated free space was increased by more than 40% with our approach. We also show that our approach trained on a synthetic dataset generalises well to real-world outdoor scenes without additional fine-tuning. Finally, we demonstrate how motion planning tasks can benefit from these denser reconstructions.
Graph Neural Network and Spatiotemporal Transformer Attention for 3D Video Object Detection from Point Clouds
Authors: Junbo Yin, Jianbing Shen, Xin Gao, David Crandall, Ruigang Yang
Subjects: Computer Vision and Pattern Recognition (cs.CV)
Abstract
Previous works for LiDAR-based 3D object detection mainly focus on the single-frame paradigm. In this paper, we propose to detect 3D objects by exploiting temporal information in multiple frames, i.e., the point cloud videos. We empirically categorize the temporal information into short-term and long-term patterns. To encode the short-term data, we present a Grid Message Passing Network (GMPNet), which considers each grid (i.e., the grouped points) as a node and constructs a k-NN graph with the neighbor grids. To update features for a grid, GMPNet iteratively collects information from its neighbors, thus mining the motion cues in grids from nearby frames. To further aggregate the long-term frames, we propose an Attentive Spatiotemporal Transformer GRU (AST-GRU), which contains a Spatial Transformer Attention (STA) module and a Temporal Transformer Attention (TTA) module. STA and TTA enhance the vanilla GRU to focus on small objects and better align the moving objects. Our overall framework supports both online and offline video object detection in point clouds. We implement our algorithm based on prevalent anchor-based and anchor-free detectors. The evaluation results on the challenging nuScenes benchmark show the superior performance of our method, achieving the 1st on the leaderboard without any bells and whistles, by the time the paper is submitted.
CENet: Toward Concise and Efficient LiDAR Semantic Segmentation for Autonomous Driving
Abstract
Accurate and fast scene understanding is one of the challenging task for autonomous driving, which requires to take full advantage of LiDAR point clouds for semantic segmentation. In this paper, we present a \textbf{concise} and \textbf{efficient} image-based semantic segmentation network, named \textbf{CENet}. In order to improve the descriptive power of learned features and reduce the computational as well as time complexity, our CENet integrates the convolution with larger kernel size instead of MLP, carefully-selected activation functions, and multiple auxiliary segmentation heads with corresponding loss functions into architecture. Quantitative and qualitative experiments conducted on publicly available benchmarks, SemanticKITTI and SemanticPOSS, demonstrate that our pipeline achieves much better mIoU and inference performance compared with state-of-the-art models. The code will be available at https://github.com/huixiancheng/CENet.
MV-FCOS3D++: Multi-View Camera-Only 4D Object Detection with Pretrained Monocular Backbones
Abstract
In this technical report, we present our solution, dubbed MV-FCOS3D++, for the Camera-Only 3D Detection track in Waymo Open Dataset Challenge 2022. For multi-view camera-only 3D detection, methods based on bird-eye-view or 3D geometric representations can leverage the stereo cues from overlapped regions between adjacent views and directly perform 3D detection without hand-crafted post-processing. However, it lacks direct semantic supervision for 2D backbones, which can be complemented by pretraining simple monocular-based detectors. Our solution is a multi-view framework for 4D detection following this paradigm. It is built upon a simple monocular detector FCOS3D++, pretrained only with object annotations of Waymo, and converts multi-view features to a 3D grid space to detect 3D objects thereon. A dual-path neck for single-frame understanding and temporal stereo matching is devised to incorporate multi-frame information. Our method finally achieves 49.75% mAPL with a single model and wins 2nd place in the WOD challenge, without any LiDAR-based depth supervision during training. The code will be released at https://github.com/Tai-Wang/Depth-from-Motion.
Keyword: loop detection
There is no result
Keyword: nerf
There is no result
Keyword: mapping
3D Lidar Reconstruction with Probabilistic Depth Completion for Robotic Navigation
Authors: Yifu Tao, Marija Popović, Yiduo Wang, Sundara Tejaswi Digumarti, Nived Chebrolu, Maurice Fallon
Abstract
Safe motion planning in robotics requires planning into space which has been verified to be free of obstacles. However, obtaining such environment representations using lidars is challenging by virtue of the sparsity of their depth measurements. We present a learning-aided 3D lidar reconstruction framework that upsamples sparse lidar depth measurements with the aid of overlapping camera images so as to generate denser reconstructions with more definitively free space than can be achieved with the raw lidar measurements alone. We use a neural network with an encoder-decoder structure to predict dense depth images along with depth uncertainty estimates which are fused using a volumetric mapping system. We conduct experiments on real-world outdoor datasets captured using a handheld sensing device and a legged robot. Using input data from a 16-beam lidar mapping a building network, our experiments showed that the amount of estimated free space was increased by more than 40% with our approach. We also show that our approach trained on a synthetic dataset generalises well to real-world outdoor scenes without additional fine-tuning. Finally, we demonstrate how motion planning tasks can benefit from these denser reconstructions.
Adaptive Asymmetric Label-guided Hashing for Multimedia Search
Authors: Yitian Long
Subjects: Information Retrieval (cs.IR); Multimedia (cs.MM)
Abstract
With the rapid growth of multimodal media data on the Web in recent years, hash learning methods as a way to achieve efficient and flexible cross-modal retrieval of massive multimedia data have received a lot of attention from the current Web resource retrieval research community. Existing supervised hashing methods simply transform label information into pairwise similarity information to guide hash learning, leading to a potential risk of semantic error in the face of multi-label data. In addition, most existing hash optimization methods solve NP-hard optimization problems by employing approximate approximation strategies based on relaxation strategies, leading to a large quantization error. In order to address above obstacles, we present a simple yet efficient Adaptive Asymmetric Label-guided Hashing, named A2LH, for Multimedia Search. Specifically, A2LH is a two-step hashing method. In the first step, we design an association representation model between the different modality representations and semantic label representation separately, and use the semantic label representation as an intermediate bridge to solve the semantic gap existing between different modalities. In addition, we present an efficient discrete optimization algorithm for solving the quantization error problem caused by relaxation-based optimization algorithms. In the second step, we leverage the generated hash codes to learn the hash mapping functions. The experimental results show that our proposed method achieves optimal performance on all compared baseline methods.
Can Deep Learning Assist Automatic Identification of Layered Pigments From XRF Data?
Authors: Bingjie (Jenny)Xu, Yunan Wu, Pengxiao Hao, Marc Vermeulen, Alicia McGeachy, Kate Smith, Katherine Eremin, Georgina Rayner, Giovanni Verri, Florian Willomitzer, Matthias Alfeld, Jack Tumblin, Aggelos Katsaggelos, Marc Walton
Subjects: Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG); Image and Video Processing (eess.IV)
Abstract
X-ray fluorescence spectroscopy (XRF) plays an important role for elemental analysis in a wide range of scientific fields, especially in cultural heritage. XRF imaging, which uses a raster scan to acquire spectra across artworks, provides the opportunity for spatial analysis of pigment distributions based on their elemental composition. However, conventional XRF-based pigment identification relies on time-consuming elemental mapping by expert interpretations of measured spectra. To reduce the reliance on manual work, recent studies have applied machine learning techniques to cluster similar XRF spectra in data analysis and to identify the most likely pigments. Nevertheless, it is still challenging for automatic pigment identification strategies to directly tackle the complex structure of real paintings, e.g. pigment mixtures and layered pigments. In addition, pixel-wise pigment identification based on XRF imaging remains an obstacle due to the high noise level compared with averaged spectra. Therefore, we developed a deep-learning-based end-to-end pigment identification framework to fully automate the pigment identification process. In particular, it offers high sensitivity to the underlying pigments and to the pigments with a low concentration, therefore enabling satisfying results in mapping the pigments based on single-pixel XRF spectrum. As case studies, we applied our framework to lab-prepared mock-up paintings and two 19th-century paintings: Paul Gauguin's Po`emes Barbares (1896) that contains layered pigments with an underlying painting, and Paul Cezanne's The Bathers (1899-1904). The pigment identification results demonstrated that our model achieved comparable results to the analysis by elemental mapping, suggesting the generalizability and stability of our model.
Learning Visual Representation from Modality-Shared Contrastive Language-Image Pre-training
Authors: Haoxuan You, Luowei Zhou, Bin Xiao, Noel Codella, Yu Cheng, Ruochen Xu, Shih-Fu Chang, Lu Yuan
Subjects: Computer Vision and Pattern Recognition (cs.CV); Computation and Language (cs.CL)
Abstract
Large-scale multi-modal contrastive pre-training has demonstrated great utility to learn transferable features for a range of downstream tasks by mapping multiple modalities into a shared embedding space. Typically, this has employed separate encoders for each modality. However, recent work suggests that transformers can support learning across multiple modalities and allow knowledge sharing. Inspired by this, we investigate a variety of Modality-Shared Contrastive Language-Image Pre-training (MS-CLIP) frameworks. More specifically, we question how many parameters of a transformer model can be shared across modalities during contrastive pre-training, and rigorously examine architectural design choices that position the proportion of parameters shared along a spectrum. In studied conditions, we observe that a mostly unified encoder for vision and language signals outperforms all other variations that separate more parameters. Additionally, we find that light-weight modality-specific parallel modules further improve performance. Experimental results show that the proposed MS-CLIP approach outperforms vanilla CLIP by up to 13\% relative in zero-shot ImageNet classification (pre-trained on YFCC-100M), while simultaneously supporting a reduction of parameters. In addition, our approach outperforms vanilla CLIP by 1.6 points in linear probing on a collection of 24 downstream vision tasks. Furthermore, we discover that sharing parameters leads to semantic concepts from different modalities being encoded more closely in the embedding space, facilitating the transferring of common semantic structure (e.g., attention patterns) from language to vision. Code is available at \href{https://github.com/Hxyou/MSCLIP}{URL}.
FRIB: Low-poisoning Rate Invisible Backdoor Attack based on Feature Repair
Abstract
During the generation of invisible backdoor attack poisoned data, the feature space transformation operation tends to cause the loss of some poisoned features and weakens the mapping relationship between source images with triggers and target labels, resulting in the need for a higher poisoning rate to achieve the corresponding backdoor attack success rate. To solve the above problems, we propose the idea of feature repair for the first time and introduce the blind watermark technique to repair the poisoned features lost during the generation of poisoned data. Under the premise of ensuring consistent labeling, we propose a low-poisoning rate invisible backdoor attack based on feature repair, named FRIB. Benefiting from the above design concept, the new method enhances the mapping relationship between the source images with triggers and the target labels, and increases the degree of misleading DNNs, thus achieving a high backdoor attack success rate with a very low poisoning rate. Ultimately, the detailed experimental results show that the goal of achieving a high success rate of backdoor attacks with a very low poisoning rate is achieved on all MNIST, CIFAR10, GTSRB, and ImageNet datasets.
Keyword: localization
RenderNet: Visual Relocalization Using Virtual Viewpoints in Large-Scale Indoor Environments
Authors: Jiahui Zhang, Shitao Tang, Kejie Qiu, Rui Huang, Chuan Fang, Le Cui, Zilong Dong, Siyu Zhu, Ping Tan
Subjects: Computer Vision and Pattern Recognition (cs.CV)
Abstract
Visual relocalization has been a widely discussed problem in 3D vision: given a pre-constructed 3D visual map, the 6 DoF (Degrees-of-Freedom) pose of a query image is estimated. Relocalization in large-scale indoor environments enables attractive applications such as augmented reality and robot navigation. However, appearance changes fast in such environments when the camera moves, which is challenging for the relocalization system. To address this problem, we propose a virtual view synthesis-based approach, RenderNet, to enrich the database and refine poses regarding this particular scenario. Instead of rendering real images which requires high-quality 3D models, we opt to directly render the needed global and local features of virtual viewpoints and apply them in the subsequent image retrieval and feature matching operations respectively. The proposed method can largely improve the performance in large-scale indoor environments, e.g., achieving an improvement of 7.1\% and 12.2\% on the Inloc dataset.
$\textbf{P$^2$A}$: A Dataset and Benchmark for Dense Action Detection from Table Tennis Match Broadcasting Videos
Authors: Jiang Bian, Qingzhong Wang, Haoyi Xiong, Jun Huang, Chen Liu, Xuhong Li, Jun Cheng, Jun Zhao, Feixiang Lu, Dejing Dou
Subjects: Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
Abstract
While deep learning has been widely used for video analytics, such as video classification and action detection, dense action detection with fast-moving subjects from sports videos is still challenging. In this work, we release yet another sports video dataset $\textbf{P$^2$A}$ for $\underline{P}$ing $\underline{P}$ong-$\underline{A}$ction detection, which consists of 2,721 video clips collected from the broadcasting videos of professional table tennis matches in World Table Tennis Championships and Olympiads. We work with a crew of table tennis professionals and referees to obtain fine-grained action labels (in 14 classes) for every ping-pong action that appeared in the dataset and formulate two sets of action detection problems - action localization and action recognition. We evaluate a number of commonly-seen action recognition (e.g., TSM, TSN, Video SwinTransformer, and Slowfast) and action localization models (e.g., BSN, BSN++, BMN, TCANet), using $\textbf{P$^2$A}$ for both problems, under various settings. These models can only achieve 48% area under the AR-AN curve for localization and 82% top-one accuracy for recognition since the ping-pong actions are dense with fast-moving subjects but broadcasting videos are with only 25 FPS. The results confirm that $\textbf{P$^2$A}$ is still a challenging task and can be used as a benchmark for action detection from videos.
Perun: Performance Version System
Authors: Tomáš Fiedor, Jiří Pavela, Adam Rogalewicz, Tomáš Vojnar
Abstract
In this paper, we present Perun: an open-source tool suite for profiling-based performance analysis. At its core, Perun maintains links between project versions and the corresponding stored performance profiles, which are then leveraged for automated detection of performance changes in new project versions. The Perun tool suite further includes multiple profilers (and is designed such that further profilers can be easily added), a performance fuzz-tester for workload generation, methods for deriving performance models, and numerous visualization methods. We demonstrate how Perun can help developers to analyze their program performance on two case studies: detection and localization of a performance degradation and generation of inputs forcing performance issues to show up.
Tracking Every Thing in the Wild
Authors: Siyuan Li, Martin Danelljan, Henghui Ding, Thomas E. Huang, Fisher Yu
Subjects: Computer Vision and Pattern Recognition (cs.CV)
Abstract
Current multi-category Multiple Object Tracking (MOT) metrics use class labels to group tracking results for per-class evaluation. Similarly, MOT methods typically only associate objects with the same class predictions. These two prevalent strategies in MOT implicitly assume that the classification performance is near-perfect. However, this is far from the case in recent large-scale MOT datasets, which contain large numbers of classes with many rare or semantically similar categories. Therefore, the resulting inaccurate classification leads to sub-optimal tracking and inadequate benchmarking of trackers. We address these issues by disentangling classification from tracking. We introduce a new metric, Track Every Thing Accuracy (TETA), breaking tracking measurement into three sub-factors: localization, association, and classification, allowing comprehensive benchmarking of tracking performance even under inaccurate classification. TETA also deals with the challenging incomplete annotation problem in large-scale tracking datasets. We further introduce a Track Every Thing tracker (TETer), that performs association using Class Exemplar Matching (CEM). Our experiments show that TETA evaluates trackers more comprehensively, and TETer achieves significant improvements on the challenging large-scale datasets BDD100K and TAO compared to the state-of-the-art.
Keyword: transformer
Multi-Attention Network for Compressed Video Referring Object Segmentation
Abstract
Referring video object segmentation aims to segment the object referred by a given language expression. Existing works typically require compressed video bitstream to be decoded to RGB frames before being segmented, which increases computation and storage requirements and ultimately slows the inference down. This may hamper its application in real-world computing resource limited scenarios, such as autonomous cars and drones. To alleviate this problem, in this paper, we explore the referring object segmentation task on compressed videos, namely on the original video data flow. Besides the inherent difficulty of the video referring object segmentation task itself, obtaining discriminative representation from compressed video is also rather challenging. To address this problem, we propose a multi-attention network which consists of dual-path dual-attention module and a query-based cross-modal Transformer module. Specifically, the dual-path dual-attention module is designed to extract effective representation from compressed data in three modalities, i.e., I-frame, Motion Vector and Residual. The query-based cross-modal Transformer firstly models the correlation between linguistic and visual modalities, and then the fused multi-modality features are used to guide object queries to generate a content-aware dynamic kernel and to predict final segmentation masks. Different from previous works, we propose to learn just one kernel, which thus removes the complicated post mask-matching procedure of existing methods. Extensive promising experimental results on three challenging datasets show the effectiveness of our method compared against several state-of-the-art methods which are proposed for processing RGB data. Source code is available at: https://github.com/DexiangHong/MANet.
Cross-Modal Causal Relational Reasoning for Event-Level Visual Question Answering
Authors: Yang Liu, Guanbin Li, Liang Lin
Subjects: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
Abstract
Existing visual question answering methods tend to capture the spurious correlations from visual and linguistic modalities, and fail to discover the true casual mechanism that facilitates reasoning truthfully based on the dominant visual evidence and the correct question intention. Additionally, the existing methods usually ignore the complex event-level understanding in multi-modal settings that requires a strong cognitive capability of causal inference to jointly model cross-modal event temporality, causality, and dynamics. In this work, we focus on event-level visual question answering from a new perspective, i.e., cross-modal causal relational reasoning, by introducing causal intervention methods to mitigate the spurious correlations and discover the true causal structures for the integration of visual and linguistic modalities. Specifically, we propose a novel event-level visual question answering framework named Cross-Modal Causal RelatIonal Reasoning (CMCIR), to achieve robust casuality-aware visual-linguistic question answering. To uncover the causal structures for visual and linguistic modalities, the novel Causality-aware Visual-Linguistic Reasoning (CVLR) module is proposed to collaboratively disentangle the visual and linguistic spurious correlations via elaborately designed front-door and back-door causal intervention modules. To discover the fine-grained interactions between linguistic semantics and spatial-temporal representations, we build a novel Spatial-Temporal Transformer (STT) that builds the multi-modal co-occurrence interactions between visual and linguistic content. Extensive experiments on large-scale event-level urban dataset SUTD-TrafficQA and three benchmark real-world datasets TGIF-QA, MSVD-QA, and MSRVTT-QA demonstrate the effectiveness of our CMCIR for discovering visual-linguistic causal structures.
Graph Neural Network and Spatiotemporal Transformer Attention for 3D Video Object Detection from Point Clouds
Authors: Junbo Yin, Jianbing Shen, Xin Gao, David Crandall, Ruigang Yang
Subjects: Computer Vision and Pattern Recognition (cs.CV)
Abstract
Previous works for LiDAR-based 3D object detection mainly focus on the single-frame paradigm. In this paper, we propose to detect 3D objects by exploiting temporal information in multiple frames, i.e., the point cloud videos. We empirically categorize the temporal information into short-term and long-term patterns. To encode the short-term data, we present a Grid Message Passing Network (GMPNet), which considers each grid (i.e., the grouped points) as a node and constructs a k-NN graph with the neighbor grids. To update features for a grid, GMPNet iteratively collects information from its neighbors, thus mining the motion cues in grids from nearby frames. To further aggregate the long-term frames, we propose an Attentive Spatiotemporal Transformer GRU (AST-GRU), which contains a Spatial Transformer Attention (STA) module and a Temporal Transformer Attention (TTA) module. STA and TTA enhance the vanilla GRU to focus on small objects and better align the moving objects. Our overall framework supports both online and offline video object detection in point clouds. We implement our algorithm based on prevalent anchor-based and anchor-free detectors. The evaluation results on the challenging nuScenes benchmark show the superior performance of our method, achieving the 1st on the leaderboard without any bells and whistles, by the time the paper is submitted.
Learning Visual Representation from Modality-Shared Contrastive Language-Image Pre-training
Authors: Haoxuan You, Luowei Zhou, Bin Xiao, Noel Codella, Yu Cheng, Ruochen Xu, Shih-Fu Chang, Lu Yuan
Subjects: Computer Vision and Pattern Recognition (cs.CV); Computation and Language (cs.CL)
Abstract
Large-scale multi-modal contrastive pre-training has demonstrated great utility to learn transferable features for a range of downstream tasks by mapping multiple modalities into a shared embedding space. Typically, this has employed separate encoders for each modality. However, recent work suggests that transformers can support learning across multiple modalities and allow knowledge sharing. Inspired by this, we investigate a variety of Modality-Shared Contrastive Language-Image Pre-training (MS-CLIP) frameworks. More specifically, we question how many parameters of a transformer model can be shared across modalities during contrastive pre-training, and rigorously examine architectural design choices that position the proportion of parameters shared along a spectrum. In studied conditions, we observe that a mostly unified encoder for vision and language signals outperforms all other variations that separate more parameters. Additionally, we find that light-weight modality-specific parallel modules further improve performance. Experimental results show that the proposed MS-CLIP approach outperforms vanilla CLIP by up to 13\% relative in zero-shot ImageNet classification (pre-trained on YFCC-100M), while simultaneously supporting a reduction of parameters. In addition, our approach outperforms vanilla CLIP by 1.6 points in linear probing on a collection of 24 downstream vision tasks. Furthermore, we discover that sharing parameters leads to semantic concepts from different modalities being encoded more closely in the embedding space, facilitating the transferring of common semantic structure (e.g., attention patterns) from language to vision. Code is available at \href{https://github.com/Hxyou/MSCLIP}{URL}.
$\textbf{P$^2$A}$: A Dataset and Benchmark for Dense Action Detection from Table Tennis Match Broadcasting Videos
Authors: Jiang Bian, Qingzhong Wang, Haoyi Xiong, Jun Huang, Chen Liu, Xuhong Li, Jun Cheng, Jun Zhao, Feixiang Lu, Dejing Dou
Subjects: Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
Abstract
While deep learning has been widely used for video analytics, such as video classification and action detection, dense action detection with fast-moving subjects from sports videos is still challenging. In this work, we release yet another sports video dataset $\textbf{P$^2$A}$ for $\underline{P}$ing $\underline{P}$ong-$\underline{A}$ction detection, which consists of 2,721 video clips collected from the broadcasting videos of professional table tennis matches in World Table Tennis Championships and Olympiads. We work with a crew of table tennis professionals and referees to obtain fine-grained action labels (in 14 classes) for every ping-pong action that appeared in the dataset and formulate two sets of action detection problems - action localization and action recognition. We evaluate a number of commonly-seen action recognition (e.g., TSM, TSN, Video SwinTransformer, and Slowfast) and action localization models (e.g., BSN, BSN++, BMN, TCANet), using $\textbf{P$^2$A}$ for both problems, under various settings. These models can only achieve 48% area under the AR-AN curve for localization and 82% top-one accuracy for recognition since the ping-pong actions are dense with fast-moving subjects but broadcasting videos are with only 25 FPS. The results confirm that $\textbf{P$^2$A}$ is still a challenging task and can be used as a benchmark for action detection from videos.
Training Effective Neural Sentence Encoders from Automatically Mined Paraphrases
Abstract
Sentence embeddings are commonly used in text clustering and semantic retrieval tasks. State-of-the-art sentence representation methods are based on artificial neural networks fine-tuned on large collections of manually labeled sentence pairs. Sufficient amount of annotated data is available for high-resource languages such as English or Chinese. In less popular languages, multilingual models have to be used, which offer lower performance. In this publication, we address this problem by proposing a method for training effective language-specific sentence encoders without manually labeled data. Our approach is to automatically construct a dataset of paraphrase pairs from sentence-aligned bilingual text corpora. We then use the collected data to fine-tune a Transformer language model with an additional recurrent pooling layer. Our sentence encoder can be trained in less than a day on a single graphics card, achieving high performance on a diverse set of sentence-level tasks. We evaluate our method on eight linguistic tasks in Polish, comparing it with the best available multilingual sentence encoders.
Bodily Behaviors in Social Interaction: Novel Annotations and State-of-the-Art Evaluation
Authors: Michal Balazia, Philipp Müller, Ákos Levente Tánczos, August von Liechtenstein, François Brémond
Subjects: Computer Vision and Pattern Recognition (cs.CV)
Abstract
Body language is an eye-catching social signal and its automatic analysis can significantly advance artificial intelligence systems to understand and actively participate in social interactions. While computer vision has made impressive progress in low-level tasks like head and body pose estimation, the detection of more subtle behaviors such as gesturing, grooming, or fumbling is not well explored. In this paper we present BBSI, the first set of annotations of complex Bodily Behaviors embedded in continuous Social Interactions in a group setting. Based on previous work in psychology, we manually annotated 26 hours of spontaneous human behavior in the MPIIGroupInteraction dataset with 15 distinct body language classes. We present comprehensive descriptive statistics on the resulting dataset as well as results of annotation quality evaluations. For automatic detection of these behaviors, we adapt the Pyramid Dilated Attention Network (PDAN), a state-of-the-art approach for human action detection. We perform experiments using four variants of spatial-temporal features as input to PDAN: Two-Stream Inflated 3D CNN, Temporal Segment Networks, Temporal Shift Module and Swin Transformer. Results are promising and indicate a great room for improvement in this difficult task. Representing a key piece in the puzzle towards automatic understanding of social behavior, BBSI is fully available to the research community.
S-Prompts Learning with Pre-trained Transformers: An Occam's Razor for Domain Incremental Learning
Authors: Yabin Wang, Zhiwu Huang, Xiaopeng Hong
Subjects: Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
Abstract
State-of-the-art deep neural networks are still struggling to address the catastrophic forgetting problem in continual learning. In this paper, we propose one simple paradigm (named as S-Prompting) and two concrete approaches to highly reduce the forgetting degree in one of the most typical continual learning scenarios, i.e., domain increment learning (DIL). The key idea of the paradigm is to learn prompts independently across domains with pre-trained transformers, avoiding the use of exemplars that commonly appear in conventional methods. This results in a win-win game where the prompting can achieve the best for each domain. The independent prompting across domains only requests one single cross-entropy loss for training and one simple K-NN operation as a domain identifier for inference. The learning paradigm derives an image prompt learning approach and a brand-new language-image prompt learning approach. Owning an excellent scalability (0.03% parameter increase per domain), the best of our approaches achieves a remarkable relative improvement (an average of about 30%) over the best of the state-of-the-art exemplar-free methods for three standard DIL tasks, and even surpasses the best of them relatively by about 6% in average when they use exemplars.
Compositional Human-Scene Interaction Synthesis with Semantic Control
Abstract
Synthesizing natural interactions between virtual humans and their 3D environments is critical for numerous applications, such as computer games and AR/VR experiences. Our goal is to synthesize humans interacting with a given 3D scene controlled by high-level semantic specifications as pairs of action categories and object instances, e.g., "sit on the chair". The key challenge of incorporating interaction semantics into the generation framework is to learn a joint representation that effectively captures heterogeneous information, including human body articulation, 3D object geometry, and the intent of the interaction. To address this challenge, we design a novel transformer-based generative model, in which the articulated 3D human body surface points and 3D objects are jointly encoded in a unified latent space, and the semantics of the interaction between the human and objects are embedded via positional encoding. Furthermore, inspired by the compositional nature of interactions that humans can simultaneously interact with multiple objects, we define interaction semantics as the composition of varying numbers of atomic action-object pairs. Our proposed generative model can naturally incorporate varying numbers of atomic interactions, which enables synthesizing compositional human-scene interactions without requiring composite interaction data. We extend the PROX dataset with interaction semantic labels and scene instance segmentation to evaluate our method and demonstrate that our method can generate realistic human-scene interactions with semantic control. Our perceptual study shows that our synthesized virtual humans can naturally interact with 3D scenes, considerably outperforming existing methods. We name our method COINS, for COmpositional INteraction Synthesis with Semantic Control. Code and data are available at https://github.com/zkf1997/COINS.
Unsupervised Domain Adaptation for Video Transformers in Action Recognition
Authors: Victor G. Turrisi da Costa, Giacomo Zara, Paolo Rota, Thiago Oliveira-Santos, Nicu Sebe, Vittorio Murino, Elisa Ricci
Subjects: Computer Vision and Pattern Recognition (cs.CV)
Abstract
Over the last few years, Unsupervised Domain Adaptation (UDA) techniques have acquired remarkable importance and popularity in computer vision. However, when compared to the extensive literature available for images, the field of videos is still relatively unexplored. On the other hand, the performance of a model in action recognition is heavily affected by domain shift. In this paper, we propose a simple and novel UDA approach for video action recognition. Our approach leverages recent advances on spatio-temporal transformers to build a robust source model that better generalises to the target domain. Furthermore, our architecture learns domain invariant features thanks to the introduction of a novel alignment loss term derived from the Information Bottleneck principle. We report results on two video action recognition benchmarks for UDA, showing state-of-the-art performance on HMDB$\leftrightarrow$UCF, as well as on Kinetics$\rightarrow$NEC-Drone, which is more challenging. This demonstrates the effectiveness of our method in handling different levels of domain shift. The source code is available at https://github.com/vturrisi/UDAVT.
Extreme compression of sentence-transformer ranker models: faster inference, longer battery life, and less storage on edge devices
Authors: Amit Chaulwar, Lukas Malik, Maciej Krajewski, Felix Reichel, Leif-Nissen Lundbæk, Michael Huth, Bartlomiej Matejczyk
Subjects: Machine Learning (cs.LG); Computation and Language (cs.CL); Information Retrieval (cs.IR)
Abstract
Modern search systems use several large ranker models with transformer architectures. These models require large computational resources and are not suitable for usage on devices with limited computational resources. Knowledge distillation is a popular compression technique that can reduce the resource needs of such models, where a large teacher model transfers knowledge to a small student model. To drastically reduce memory requirements and energy consumption, we propose two extensions for a popular sentence-transformer distillation procedure: generation of an optimal size vocabulary and dimensionality reduction of the embedding dimension of teachers prior to distillation. We evaluate these extensions on two different types of ranker models. This results in extremely compressed student models whose analysis on a test dataset shows the significance and utility of our proposed extensions.
TransFiner: A Full-Scale Refinement Approach for Multiple Object Tracking
Authors: Bin Sun, Jiale Cao
Subjects: Computer Vision and Pattern Recognition (cs.CV)
Abstract
Multiple object tracking (MOT) is the task containing detection and association. Plenty of trackers have achieved competitive performance. Unfortunately, for the lack of informative exchange on these subtasks, they are often biased toward one of the two and remain underperforming in complex scenarios, such as the expected false negatives and mistaken trajectories of targets when passing each other. In this paper, we propose TransFiner, a transformer-based post-refinement approach for MOT. It is a generic attachment framework that leverages the images and tracking results (locations and class predictions) from the original tracker as inputs, which are then used to launch TransFiner powerfully. Moreover, TransFiner depends on query pairs, which produce pairs of detection and motion through the fusion decoder and achieve comprehensive tracking improvement. We also provide targeted refinement by labeling query pairs according to different refinement levels. Experiments show that our design is effective, on the MOT17 benchmark, we elevate the CenterTrack from 67.8% MOTA and 64.7% IDF1 to 71.5% MOTA and 66.8% IDF1.
Group DETR: Fast Training Convergence with Decoupled One-to-Many Label Assignment
Authors: Qiang Chen, Xiaokang Chen, Gang Zeng, Jingdong Wang
Subjects: Computer Vision and Pattern Recognition (cs.CV)
Abstract
Detection Transformer (DETR) relies on One-to-One label assignment, i.e., assigning one ground-truth (gt) object to only one positive object query, for end-to-end object detection and lacks the capability of exploiting multiple positive queries. We present a novel DETR training approach, named {\em Group DETR}, to support multiple positive queries. To be specific, we decouple the positives into multiple independent groups and keep only one positive per gt object in each group. We make simple modifications during training: (i) adopt $K$ groups of object queries; (ii) conduct decoder self-attention on each group of object queries with the same parameters; (iii) perform One-to-One label assignment for each group, leading to $K$ positive object queries for each gt object. In inference, we only use one group of object queries, making no modifications to both architecture and processes. We validate the effectiveness of the proposed approach on DETR variants, including Conditional DETR, DAB-DETR, DN-DETR, and DINO.
Keyword: autonomous driving
$p$-DkNN: Out-of-Distribution Detection Through Statistical Testing of Deep Representations
Authors: Adam Dziedzic, Stephan Rabanser, Mohammad Yaghini, Armin Ale, Murat A. Erdogdu, Nicolas Papernot
Abstract
The lack of well-calibrated confidence estimates makes neural networks inadequate in safety-critical domains such as autonomous driving or healthcare. In these settings, having the ability to abstain from making a prediction on out-of-distribution (OOD) data can be as important as correctly classifying in-distribution data. We introduce $p$-DkNN, a novel inference procedure that takes a trained deep neural network and analyzes the similarity structures of its intermediate hidden representations to compute $p$-values associated with the end-to-end model prediction. The intuition is that statistical tests performed on latent representations can serve not only as a classifier, but also offer a statistically well-founded estimation of uncertainty. $p$-DkNN is scalable and leverages the composition of representations learned by hidden layers, which makes deep representation learning successful. Our theoretical analysis builds on Neyman-Pearson classification and connects it to recent advances in selective classification (reject option). We demonstrate advantageous trade-offs between abstaining from predicting on OOD inputs and maintaining high accuracy on in-distribution inputs. We find that $p$-DkNN forces adaptive attackers crafting adversarial examples, a form of worst-case OOD inputs, to introduce semantically meaningful changes to the inputs.
Inter-Frame Compression for Dynamic Point Cloud Geometry Coding
Authors: Anique Akhtar, Zhu Li, Geert Van der Auwera
Subjects: Computer Vision and Pattern Recognition (cs.CV); Multimedia (cs.MM); Image and Video Processing (eess.IV)
Abstract
Efficient point cloud compression is essential for applications like virtual and mixed reality, autonomous driving, and cultural heritage. In this paper, we propose a deep learning-based inter-frame encoding scheme for dynamic point cloud geometry compression. We propose a lossy geometry compression scheme that predicts the latent representation of the current frame using the previous frame by employing a novel prediction network. Our proposed network utilizes sparse convolutions with hierarchical multiscale 3D feature learning to encode the current frame using the previous frame. We employ convolution on target coordinates to map the latent representation of the previous frame to the downsampled coordinates of the current frame to predict the current frame's feature embedding. Our framework transmits the residual of the predicted features and the actual features by compressing them using a learned probabilistic factorized entropy model. At the receiver, the decoder hierarchically reconstructs the current frame by progressively rescaling the feature embedding. We compared our model to the state-of-the-art Video-based Point Cloud Compression (V-PCC) and Geometry-based Point Cloud Compression (G-PCC) schemes standardized by the Moving Picture Experts Group (MPEG). Our method achieves more than 91% BD-Rate Bjontegaard Delta Rate) reduction against G-PCC, more than 62% BD-Rate reduction against V-PCC intra-frame encoding mode, and more than 52% BD-Rate savings against V-PCC P-frame-based inter-frame encoding mode using HEVC.
Semi-supervised 3D Object Detection with Proficient Teachers
Authors: Junbo Yin, Jin Fang, Dingfu Zhou, Liangjun Zhang, Cheng-Zhong Xu, Jianbing Shen, Wenguan Wang
Subjects: Computer Vision and Pattern Recognition (cs.CV)
Abstract
Dominated point cloud-based 3D object detectors in autonomous driving scenarios rely heavily on the huge amount of accurately labeled samples, however, 3D annotation in the point cloud is extremely tedious, expensive and time-consuming. To reduce the dependence on large supervision, semi-supervised learning (SSL) based approaches have been proposed. The Pseudo-Labeling methodology is commonly used for SSL frameworks, however, the low-quality predictions from the teacher model have seriously limited its performance. In this work, we propose a new Pseudo-Labeling framework for semi-supervised 3D object detection, by enhancing the teacher model to a proficient one with several necessary designs. First, to improve the recall of pseudo labels, a Spatialtemporal Ensemble (STE) module is proposed to generate sufficient seed boxes. Second, to improve the precision of recalled boxes, a Clusteringbased Box Voting (CBV) module is designed to get aggregated votes from the clustered seed boxes. This also eliminates the necessity of sophisticated thresholds to select pseudo labels. Furthermore, to reduce the negative influence of wrongly pseudo-labeled samples during the training, a soft supervision signal is proposed by considering Box-wise Contrastive Learning (BCL). The effectiveness of our model is verified on both ONCE and Waymo datasets. For example, on ONCE, our approach significantly improves the baseline by 9.51 mAP. Moreover, with half annotations, our model outperforms the oracle model with full annotations on Waymo.
CENet: Toward Concise and Efficient LiDAR Semantic Segmentation for Autonomous Driving
Abstract
Accurate and fast scene understanding is one of the challenging task for autonomous driving, which requires to take full advantage of LiDAR point clouds for semantic segmentation. In this paper, we present a \textbf{concise} and \textbf{efficient} image-based semantic segmentation network, named \textbf{CENet}. In order to improve the descriptive power of learned features and reduce the computational as well as time complexity, our CENet integrates the convolution with larger kernel size instead of MLP, carefully-selected activation functions, and multiple auxiliary segmentation heads with corresponding loss functions into architecture. Quantitative and qualitative experiments conducted on publicly available benchmarks, SemanticKITTI and SemanticPOSS, demonstrate that our pipeline achieves much better mIoU and inference performance compared with state-of-the-art models. The code will be available at https://github.com/huixiancheng/CENet.
New submissions for Wed, 27 Jul 22
Keyword: SLAM
There is no result
Keyword: odometry
There is no result
Keyword: livox
There is no result
Keyword: loam
There is no result
Keyword: lidar
3D Lidar Reconstruction with Probabilistic Depth Completion for Robotic Navigation
Graph Neural Network and Spatiotemporal Transformer Attention for 3D Video Object Detection from Point Clouds
CENet: Toward Concise and Efficient LiDAR Semantic Segmentation for Autonomous Driving
MV-FCOS3D++: Multi-View Camera-Only 4D Object Detection with Pretrained Monocular Backbones
Keyword: loop detection
There is no result
Keyword: nerf
There is no result
Keyword: mapping
3D Lidar Reconstruction with Probabilistic Depth Completion for Robotic Navigation
Adaptive Asymmetric Label-guided Hashing for Multimedia Search
Can Deep Learning Assist Automatic Identification of Layered Pigments From XRF Data?
Learning Visual Representation from Modality-Shared Contrastive Language-Image Pre-training
FRIB: Low-poisoning Rate Invisible Backdoor Attack based on Feature Repair
Keyword: localization
RenderNet: Visual Relocalization Using Virtual Viewpoints in Large-Scale Indoor Environments
$\textbf{P$^2$A}$: A Dataset and Benchmark for Dense Action Detection from Table Tennis Match Broadcasting Videos
Perun: Performance Version System
Tracking Every Thing in the Wild
Keyword: transformer
Multi-Attention Network for Compressed Video Referring Object Segmentation
Cross-Modal Causal Relational Reasoning for Event-Level Visual Question Answering
Graph Neural Network and Spatiotemporal Transformer Attention for 3D Video Object Detection from Point Clouds
Learning Visual Representation from Modality-Shared Contrastive Language-Image Pre-training
$\textbf{P$^2$A}$: A Dataset and Benchmark for Dense Action Detection from Table Tennis Match Broadcasting Videos
Training Effective Neural Sentence Encoders from Automatically Mined Paraphrases
Bodily Behaviors in Social Interaction: Novel Annotations and State-of-the-Art Evaluation
S-Prompts Learning with Pre-trained Transformers: An Occam's Razor for Domain Incremental Learning
Compositional Human-Scene Interaction Synthesis with Semantic Control
Unsupervised Domain Adaptation for Video Transformers in Action Recognition
Extreme compression of sentence-transformer ranker models: faster inference, longer battery life, and less storage on edge devices
TransFiner: A Full-Scale Refinement Approach for Multiple Object Tracking
Group DETR: Fast Training Convergence with Decoupled One-to-Many Label Assignment
Keyword: autonomous driving
$p$-DkNN: Out-of-Distribution Detection Through Statistical Testing of Deep Representations
Inter-Frame Compression for Dynamic Point Cloud Geometry Coding
Semi-supervised 3D Object Detection with Proficient Teachers
CENet: Toward Concise and Efficient LiDAR Semantic Segmentation for Autonomous Driving