We present HOI4D, a large-scale 4D egocentric dataset with rich annotations, to catalyze the research of category-level human-object interaction. HOI4D consists of 2.4M RGB-D egocentric video frames over 4000 sequences collected by 4 participants interacting with 800 different object instances from 16 categories over 610 different indoor rooms. Frame-wise annotations for panoptic segmentation, motion segmentation, 3D hand pose, category-level object pose and hand action have also been provided, together with reconstructed object meshes and scene point clouds. With HOI4D, we establish three benchmarking tasks to promote category-level HOI from 4D visual signals including semantic segmentation of 4D dynamic point cloud sequences, category-level object pose tracking, and egocentric action segmentation with diverse interaction targets. In-depth analysis shows HOI4D poses great challenges to existing methods and produces great research opportunities.
This work proposes a novel pose estimation model for object categories that can be effectively transferred to previously unseen environments. The deep convolutional network models (CNN) for pose estimation are typically trained and evaluated on datasets specifically curated for object detection, pose estimation, or 3D reconstruction, which requires large amounts of training data. In this work, we propose a model for pose estimation that can be trained with small amount of data and is built on the top of generic mid-level representations \cite{taskonomy2018} (e.g. surface normal estimation and re-shading). These representations are trained on a large dataset without requiring pose and object annotations. Later on, the predictions are refined with a small CNN neural network that exploits object masks and silhouette retrieval. The presented approach achieves superior performance on the Pix3D dataset \cite{pix3d} and shows nearly 35\% improvement over the existing models when only 25\% of the training data is available. We show that the approach is favorable when it comes to generalization and transfer to novel environments. Towards this end, we introduce a new pose estimation benchmark for commonly encountered furniture categories on challenging Active Vision Dataset \cite{Ammirato2017ADF} and evaluated the models trained on the Pix3D dataset.
This paper proposes a universal framework, called OVE6D, for model-based 6D object pose estimation from a single depth image and a target object mask. Our model is trained using purely synthetic data rendered from ShapeNet, and, unlike most of the existing methods, it generalizes well on new real-world objects without any fine-tuning. We achieve this by decomposing the 6D pose into viewpoint, in-plane rotation around the camera optical axis and translation, and introducing novel lightweight modules for estimating each component in a cascaded manner. The resulting network contains less than 4M parameters while demonstrating excellent performance on the challenging T-LESS and Occluded LINEMOD datasets without any dataset-specific training. We show that OVE6D outperforms some contemporary deep learning-based pose estimation methods specifically trained for individual objects or datasets with real-world training data. The implementation and the pre-trained model will be made publicly available.
We propose a method for 3D object reconstruction and 6D-pose estimation from 2D images that uses knowledge about object shape as the primary key. In the proposed pipeline, recognition and labeling of objects in 2D images deliver 2D segment silhouettes that are compared with the 2D silhouettes of projections obtained from various views of a 3D model representing the recognized object class. By computing transformation parameters directly from the 2D images, the number of free parameters required during the registration process is reduced, making the approach feasible. Furthermore, 3D transformations and projective geometry are employed to arrive at a full 3D reconstruction of the object in camera space using a calibrated set up. Inclusion of a second camera allows resolving remaining ambiguities. The method is quantitatively evaluated using synthetic data and tested with real data, and additional results for the well-known Linemod data set are shown. In robot experiments, successful grasping of objects demonstrates its usability in real-world environments, and, where possible, a comparison with other methods is provided. The method is applicable to scenarios where 3D object models, e.g., CAD-models or point clouds, are available and precise pixel-wise segmentation maps of 2D images can be obtained. Different from other methods, the method does not use 3D depth for training, widening the domain of application.
In this paper, we propose a learning-based moving-object tracking method utilizing our newly developed LiDAR sensor, Frequency Modulated Continuous Wave (FMCW) LiDAR. Compared with most existing commercial LiDAR sensors, our FMCW LiDAR can provide additional Doppler velocity information to each 3D point of the point clouds. Benefiting from this, we can generate instance labels as ground truth in a semi-automatic manner. Given the labels, we propose a contrastive learning framework, which pulls together the features from the same instance in embedding space and pushes apart the features from different instances, to improve the tracking quality. Extensive experiments are conducted on our recorded driving data, and the results show that our method outperforms the baseline methods by a large margin.
Camera and LiDAR sensor modalities provide complementary appearance and geometric information useful for detecting 3D objects for autonomous vehicle applications. However, current end-to-end fusion methods are challenging to train and underperform state-of-the-art LiDAR-only detectors. Sequential fusion methods suffer from a limited number of pixel and point correspondences due to point cloud sparsity, or their performance is strictly capped by the detections of one of the modalities. Our proposed solution, Dense Voxel Fusion (DVF) is a sequential fusion method that generates multi-scale dense voxel feature representations, improving expressiveness in low point density regions. To enhance multi-modal learning, we train directly with projected ground truth 3D bounding box labels, avoiding noisy, detector-specific 2D predictions. Both DVF and the multi-modal training approach can be applied to any voxel-based LiDAR backbone. DVF ranks 3rd among published fusion methods on KITTI 3D car detection benchmark without introducing additional trainable parameters, nor requiring stereo images or dense depth labels. In addition, DVF significantly improves 3D vehicle detection performance of voxel-based methods on the Waymo Open Dataset.
Object identification is one of the most fundamental and difficult issues in computer vision. It aims to discover object instances in real pictures from a huge number of established categories. In recent years, deep learning-based object detection techniques that developed from computer vision have grabbed the public's interest. Object recognition methods based on deep learning frameworks have quickly become a popular way to interpret moving images acquired by various sensors. Due to its vast variety of applications for various computer vision tasks such as activity or event detection, content-based image retrieval, and scene understanding, academics have spent decades attempting to solve this problem. With this goal in mind, a unique deep learning classification technique is used to create an autonomous object detecting system. The noise destruction and normalising operations, which are carried out using gaussian filter and contrast normalisation techniques, respectively, are the first steps in the study activity. The pre-processed picture is next subjected to entropy-based segmentation algorithms, which separate the image's significant areas in order to distinguish between distinct occurrences. The classification challenge is completed by the suggested Hybrid Optimized Dense Convolutional Neural Network (HODCNN). The major goal of this framework is to aid in the precise recognition of distinct items from the gathered input frames. The suggested system's performance is assessed by comparing it to existing machine learning and deep learning methodologies. The experimental findings reveal that the suggested framework has a detection accuracy of 0.9864, which is greater than current techniques. As a result, the suggested object detection model outperforms other current methods.
This paper proposes a self-supervised monocular image-to-depth prediction framework that is trained with an end-to-end photometric loss that handles not only 6-DOF camera motion but also 6-DOF moving object instances. Self-supervision is performed by warping the images across a video sequence using depth and scene motion including object instances. One novelty of the proposed method is the use of the multi-head attention of the transformer network that matches moving objects across time and models their interaction and dynamics. This enables accurate and robust pose estimation for each object instance. Most image-to-depth predication frameworks make the assumption of rigid scenes, which largely degrades their performance with respect to dynamic objects. Only a few SOTA papers have accounted for dynamic objects. The proposed method is shown to outperform these methods on standard benchmarks and the impact of the dynamic motion on these benchmarks is exposed. Furthermore, the proposed image-to-depth prediction framework is also shown to be competitive with SOTA video-to-depth prediction frameworks.
Inspired by the human ability to perform complex manipulation in the complete absence of vision (like retrieving an object from a pocket), the robotic manipulation field is motivated to develop new methods for tactile-based object interaction. However, tactile sensing presents the challenge of being an active sensing modality: a touch sensor provides sparse, local data, and must be used in conjunction with effective exploration strategies in order to collect information. In this work, we focus on the process of guiding tactile exploration, and its interplay with task-related decision making. We propose TANDEM (TActile exploration aNd DEcision Making), an architecture to learn efficient exploration strategies in conjunction with decision making. Our approach is based on separate but co-trained modules for exploration and discrimination. We demonstrate this method on a tactile object recognition task, where a robot equipped with a touch sensor must explore and identify an object from a known set based on binary contact signals alone. TANDEM achieves higher accuracy with fewer actions than alternative methods and is also shown to be more robust to sensor noise.
Dynamic scene graph generation from a video is challenging due to the temporal dynamics of the scene and the inherent temporal fluctuations of predictions. We hypothesize that capturing long-term temporal dependencies is the key to effective generation of dynamic scene graphs. We propose to learn the long-term dependencies in a video by capturing the object-level consistency and inter-object relationship dynamics over object-level long-term tracklets using transformers. Experimental results demonstrate that our Dynamic Scene Graph Detection Transformer (DSG-DETR) outperforms state-of-the-art methods by a significant margin on the benchmark dataset Action Genome. Our ablation studies validate the effectiveness of each component of the proposed approach. The source code is available at https://github.com/Shengyu-Feng/DSG-DETR.
Grasping made impressive progress during the last few years thanks to deep learning. However, there are many objects for which it is not possible to choose a grasp by only looking at an RGB-D image, might it be for physical reasons (e.g., a hammer with uneven mass distribution) or task constraints (e.g., food that should not be spoiled). In such situations, the preferences of experts need to be taken into account. In this paper, we introduce a data-efficient grasping pipeline (Latent Space GP Selector -- LGPS) that learns grasp preferences with only a few labels per object (typically 1 to 4) and generalizes to new views of this object. Our pipeline is based on learning a latent space of grasps with a dataset generated with any state-of-the-art grasp generator (e.g., Dex-Net). This latent space is then used as a low-dimensional input for a Gaussian process classifier that selects the preferred grasp among those proposed by the generator. The results show that our method outperforms both GR-ConvNet and GG-CNN (two state-of-the-art methods that are also based on labeled grasps) on the Cornell dataset, especially when only a few labels are used: only 80 labels are enough to correctly choose 80% of the grasps (885 scenes, 244 objects). Results are similar on our dataset (91 scenes, 28 objects).
An object detection pipeline comprises a camera that captures the scene and an object detector that processes these images. The quality of the images directly affects the performance of the object detector. Many works nowadays focus either on improving the image quality or improving the object detection models independently, but neglect the importance of joint optimization of the two subsystems. The goal of this paper is to tune the detection throughput and accuracy of existing object detectors in the remote sensing scenario by focusing on optimizing the input images tailored to the object detector. To achieve this, we empirically analyze the influence of two selected camera calibration parameters (camera distortion correction and gamma correction) and five image parameters (quantization, compression, resolution, color model, additional channels) for these applications. For our experiments, we utilize three UAV data sets from different domains and a mixture of large and small state-of-the-art object detector models to provide an extensive evaluation of the influence of the pipeline parameters. Finally, we realize an object detection pipeline prototype on an embedded platform for an UAV and give a best practice recommendation for building object detection pipelines based on our findings. We show that not all parameters have an equal impact on detection accuracy and data throughput, and that by using a suitable compromise between parameters we are able to achieve higher detection accuracy for lightweight object detection models, while keeping the same data throughput.
As technology advances, the need for safe, efficient, and collaborative human-robot-teams has become increasingly important. One of the most fundamental collaborative tasks in any setting is the object handover. Human-to-robot handovers can take either of two approaches: (1) direct hand-to-hand or (2) indirect hand-to-placement-to-pick-up. The latter approach ensures minimal contact between the human and robot but can also result in increased idle time due to having to wait for the object to first be placed down on a surface. To minimize such idle time, the robot must preemptively predict the human intent of where the object will be placed. Furthermore, for the robot to preemptively act in any sort of productive manner, predictions and motion planning must occur in real-time. We introduce a novel prediction-planning pipeline that allows the robot to preemptively move towards the human agent's intended placement location using gaze and gestures as model inputs. In this paper, we investigate the performance and drawbacks of our early intent predictor-planner as well as the practical benefits of using such a pipeline through a human-robot case study.
Robots operating in human-centered environments should have the ability to understand how objects function: what can be done with each object, where this interaction may occur, and how the object is used to achieve a goal. To this end, we propose a novel approach that extracts a self-supervised visual affordance model from human teleoperated play data and leverages it to enable efficient policy learning and motion planning. We combine model-based planning with model-free deep reinforcement learning (RL) to learn policies that favor the same object regions favored by people, while requiring minimal robot interactions with the environment. We evaluate our algorithm, Visual Affordance-guided Policy Optimization (VAPO), with both diverse simulation manipulation tasks and real world robot tidy-up experiments to demonstrate the effectiveness of our affordance-guided policies. We find that our policies train 4x faster than the baselines and generalize better to novel objects because our visual affordance model can anticipate their affordance regions.
Understanding what objects could furnish for humans-namely, learning object affordance-is the crux to bridge perception and action. In the vision community, prior work primarily focuses on learning object affordance with dense (e.g., at a per-pixel level) supervision. In stark contrast, we humans learn the object affordance without dense labels. As such, the fundamental question to devise a computational model is: What is the natural way to learn the object affordance from visual appearance and geometry with humanlike sparse supervision? In this work, we present a new task of part-level affordance discovery (PartAfford): Given only the affordance labels per object, the machine is tasked to (i) decompose 3D shapes into parts and (ii) discover how each part of the object corresponds to a certain affordance category. We propose a novel learning framework for PartAfford, which discovers part-level representations by leveraging only the affordance set supervision and geometric primitive regularization, without dense supervision. The proposed approach consists of two main components: (i) an abstraction encoder with slot attention for unsupervised clustering and abstraction, and (ii) an affordance decoder with branches for part reconstruction, affordance prediction, and cuboidal primitive regularization. To learn and evaluate PartAfford, we construct a part-level, cross-category 3D object affordance dataset, annotated with 24 affordance categories shared among >25, 000 objects. We demonstrate that our method enables both the abstraction of 3D objects and part-level affordance discovery, with generalizability to difficult and cross-category examples. Further ablations reveal the contribution of each component.
Object perception plays a fundamental role in Cooperative Driving Automation (CDA) which is regarded as a revolutionary promoter for the next-generation transportation systems. However, the vehicle-based perception may suffer from the limited sensing range and occlusion as well as low penetration rates in connectivity. In this paper, we propose Cyber Mobility Mirror (CMM), a next-generation real-time traffic surveillance system for 3D object perception and reconstruction, to explore the potential of roadside sensors for enabling CDA in the real world. The CMM system consists of six main components: 1) the data pre-processor to retrieve and preprocess the raw data; 2) the roadside 3D object detector to generate 3D detection results; 3) the multi-object tracker to identify detected objects; 4) the global locator to map positioning information from the LiDAR coordinate to geographic coordinate using coordinate transformation; 5) the cloud-based communicator to transmit perception information from roadside sensors to equipped vehicles, and 6) the onboard advisor to reconstruct and display the real-time traffic conditions via Graphical User Interface (GUI). In this study, a field-operational system is deployed at a real-world intersection, University Avenue and Iowa Avenue in Riverside, California to assess the feasibility and performance of our CMM system. Results from field tests demonstrate that our CMM prototype system can provide satisfactory perception performance with 96.99% precision and 83.62% recall. High-fidelity real-time traffic conditions (at the object level) can be geo-localized with an average error of 0.14m and displayed on the GUI of the equipped vehicle with a frequency of 3-4 Hz.
Current state-of-the-art one-stage object detectors are limited by treating each image region separately without considering possible relations of the objects. This causes dependency solely on high-quality convolutional feature representations for detecting objects successfully. However, this may not be possible sometimes due to some challenging conditions. In this paper, the usage of reasoning features on one-stage object detection is analyzed. We attempted different architectures that reason the relations of the image regions by using self-attention. YOLOv3-Reasoner2 model spatially and semantically enhances features in the reasoning layer and fuses them with the original convolutional features to improve performance. The YOLOv3-Reasoner2 model achieves around 2.5% absolute improvement with respect to baseline YOLOv3 on COCO in terms of mAP while still running in real-time.
Automated product recognition in retail stores is an important real-world application in the domain of Computer Vision and Pattern Recognition. In this paper, we consider the problem of automatically identifying the classes of the products placed on racks in retail stores from an image of the rack and information about the query/product images. We improve upon the existing approaches in terms of effectiveness and memory requirement by developing a two-stage object detection and recognition pipeline comprising of a Faster-RCNN-based object localizer that detects the object regions in the rack image and a ResNet-18-based image encoder that classifies the detected regions into the appropriate classes. Each of the models is fine-tuned using appropriate data sets for better prediction and data augmentation is performed on each query image to prepare an extensive gallery set for fine-tuning the ResNet-18-based product recognition model. This encoder is trained using a triplet loss function following the strategy of online-hard-negative-mining for improved prediction. The proposed models are lightweight and can be connected in an end-to-end manner during deployment for automatically identifying each product object placed in a rack image. Extensive experiments using Grozi-32k and GP-180 data sets verify the effectiveness of the proposed model.
Hard example mining methods generally improve the performance of the object detectors, which suffer from imbalanced training sets. In this work, two existing hard example mining approaches (LRM and focal loss, FL) are adapted and combined in a state-of-the-art real-time object detector, YOLOv5. The effectiveness of the proposed approach for improving the performance on hard examples is extensively evaluated. The proposed method increases mAP by 3% compared to using the original loss function and around 1-2% compared to using the hard-mining methods (LRM or FL) individually on 2021 Anti-UAV Challenge Dataset.
The semantic gap is defined as the difference between the linguistic representations of the same concept, which usually leads to misunderstanding between individuals with different knowledge backgrounds. Since linguistically annotated images are extensively used for training machine learning models, semantic gap problem (SGP) also results in inevitable bias on image annotations and further leads to poor performance on current computer vision tasks. To address this problem, we propose a novel unsupervised method to build visual semantics aware object hierarchy, aiming to get a classification model by learning from pure-visual information and to dissipate the bias of linguistic representations caused by SGP. Our intuition in this paper comes from real-world knowledge representation where concepts are hierarchically organized, and each concept can be described by a set of features rather than a linguistic annotation, namely visual semantic. The evaluation consists of two parts, firstly we apply the constructed hierarchy on the object recognition task and then we compare our visual hierarchy and existing lexical hierarchies to show the validity of our method. The preliminary results reveal the efficiency and potential of our proposed method.