Data augmentation plays an important role in visual-based 3D object detection. Existing detectors typically employ image/BEV-level data augmentation techniques, failing to utilize flexible object-level augmentations because of 2D-3D inconsistencies. This limitation hinders us from increasing the diversity of training data. To alleviate this issue, we propose an object-level data augmentation approach that incorporates scene reconstruction and neural scene rendering. Specifically, we reconstruct the scene and objects by extracting image features from sequences and aligning them with associated LiDAR point clouds. This approach is intended to conduct the editing process within a 3D space, allowing for flexible object manipulation. Additionally, we introduce a neural scene renderer to project the edited 3D scene onto a specified camera plane and render it onto a 2D image. Combined with scene reconstruction, it overcomes the challenges stemming from 2D/3D inconsistencies, enabling the generation of object-level augmented images with corresponding labels for model training. To validate the proposed method, we apply our method to various multi-camera 3D object detectors, consistently boosting the performance.

Multi-camera perception tasks have gained significant attention in the field of autonomous driving. However, existing frameworks based on Lift-Splat-Shoot (LSS) in the multi-camera setting cannot produce suitable dense 3D features due to the projection nature and uncontrollable densification process. To resolve this problem, we propose to regulate intermediate dense 3D features with the help of volume rendering. Specifically, we employ volume rendering to process the dense 3D features to obtain corresponding 2D features (e.g., depth maps, semantic maps), which are supervised by associated labels in the training. This manner regulates the generation of dense 3D features on the feature level, providing appropriate dense and unified features for multiple perception tasks. Therefore, our approach is termed Vampire, stands for ``Volume rendering As Multi-camera Perception Intermediate feature REgulator''. Experimental results on the Occ3D and nuScenes datasets demonstrate that Vampire facilitates fine-grained and appropriate extraction of dense 3D features, and is competitive with existing SOTA methods across diverse downstream perception tasks like 3D occupancy prediction, LiDAR segmentation and 3D objection detection, while utilizing moderate GPU resources. We provide a video demonstration in the supplementary materials and Codes are available at github.com/cskkxjk/Vampire.

Monocular 3D detection is a challenging task due to the lack of accurate 3D information. Existing approaches typically rely on geometry constraints and dense depth esti-mates to facilitate the learning, but often fail to fully ex-ploit the benefits of three-dimensional feature extraction in frustum and 3D space. In this paper, we propose Occu- pancyM3D, a method of learning occupancy for monocu-lar 3D detection. It directly learns occupancy in frustum and 3D space, leading to more discriminative and informative 3D features and representations. Specifically, by using synchronized raw sparse LiDAR point clouds, we define the space status and generate voxel-based occupancy labels. We formulate occupancy prediction as a simple classification problem and design associated occupancy losses. Re-sulting occupancy estimates are employed to enhance orig-inal frustum/3D features. As a result, experiments on KITTI and Waymo open datasets demonstrate that the proposed method achieves a new state of the art and surpasses other methods by a significant margin.

In the field of monocular 3D detection, it is common practice to utilize scene geometric clues to enhance the detector’s performance. However, many existing works adopt these clues explicitly such as estimating a depth map and back-projecting it into 3D space. This explicit methodology induces sparsity in 3D representations due to the increased dimensionality from 2D to 3D, and leads to substantial information loss, especially for distant and occluded objects. To alleviate this issue, we propose MonoNeRD, a novel detection framework that can infer dense 3D geometry and occupancy. Specifically, we model scenes with Signed Distance Functions (SDF), facilitating the production of dense 3D representations. We treat these representations as Neural Radiance Fields (NeRF) and then employ volume rendering to recover RGB images and depth maps. To the best of our knowledge, this work is the first to introduce volume rendering for M3D, and demonstrates the potential of implicit reconstruction for image-based 3D perception. Extensive experiments conducted on the KITTI-3D benchmark and Waymo Open Dataset demonstrate the effectiveness of MonoNeRD. Codes are available at https://github.com/cskkxjk/MonoNeRD.

Arm motion timing is critical during basketball shooting. This study used a body-worn, sensorized basketball sleeve to identify arm motion timing characteristics during basketball free throw and jump shot shooting for trained and novice shooters. Current basketball shooting research has typically focused on arm kinematic angles, while shot timing has received comparatively less attention. An experiment was conducted to compare arm motion timing between trained and novice shooters while shooting free throws, and a second experiment compared arm motion timing between free throws and jump shots by trained shooters. Trained shooters shot free throws significantly faster than novice shooters, and trained shooters shot jump shots significantly faster than free throws at the same distance from the basket. Knowledge of arm motion timing characteristics from this study could enable future training for improved shooter accuracy.

Gait retraining is an important rehabilitation method for re-establishing health gait patterns resulting from disease or injury. Optical marker-based motion capture systems are effective for sensing but aren't used widely, due to cost and lack of portability. Moreover, to perform gait retraining, feedback is needed in addition to sensing. This paper presents a wearable sensing and haptic feedback research platform for gait retraining. The platform contains eight distributed nodes (Dots) and a central control unit (Hub) that wirelessly connects to the Dots. Each Dot provides 9-axis inertial sensing and can be configured for sensing or/and providing vibrotactile feedback according to movement training requirements. The Hub receives the sensor data, performs algorithm computation and distributes feedback commands based on the feedback strategy. A foot progression angle (FPA) gait retraining task was performed by six healthy older adults. Participants used the wearable system to learn toe-in gait (foot pointing more inward) and toe-out gait (foot pointing more outward) by adjusting their FPA based on haptic cues to fall within the no feedback zone, i.e. the desired range of acceptable FPAs. After gait retraining, FPA during toe-in gait (1.8±5.6 deg) was significantly higher than during baseline walking (-4.3±5.1 deg) (p<;0.01) and during toe-out gait (-9.9±3.2 deg) (p<;0.01). The no feedback zone was easily found by participants as the percentage of time with no feedback for toe-in gait was 68.3%, and for toe-out gait it was 89.4%. This work demonstrates that the wearable system can be an effective gait retraining research platform.