Building 3D maps of the environment is central to robot navigation, planning, and interaction with objects in a scene. Most existing approaches that integrate semantic concepts with 3D maps largely remain confined to the closed-set setting: they can only reason about a finite set of concepts, pre-defined at training time. Further, these maps can only be queried using class labels, or in recent work, using text prompts. We address both these issues with ConceptFusion, a scene representation that is (1) fundamentally open-set, enabling reasoning beyond a closed set of concepts and (ii) inherently multimodal, enabling a diverse range of possible queries to the 3D map, from language, to images, to audio, to 3D geometry, all working in concert. ConceptFusion leverages the open-set capabilities of today's foundation models pre-trained on internet-scale data to reason about concepts across modalities such as natural language, images, and audio. We demonstrate that pixel-aligned open-set features can be fused into 3D maps via traditional SLAM and multi-view fusion approaches. This enables effective zero-shot spatial reasoning, not needing any additional training or finetuning, and retains long-tailed concepts better than supervised approaches, outperforming them by more than 40% margin on 3D IoU. We extensively evaluate ConceptFusion on a number of real-world datasets, simulated home environments, a real-world tabletop manipulation task, and an autonomous driving platform. We showcase new avenues for blending foundation models with 3D open-set multimodal mapping. For more information, visit our project page https://concept-fusion.github.io or watch our 5-minute explainer video https://www.youtube.com/watch?v=rkXgws8fiDs
Few-shot semantic segmentation is the task of learning to locate each pixel of the novel class in the query image with only a few annotated support images. The current correlation-based methods construct pair-wise feature correlations to establish the many-to-many matching because the typical prototype-based approaches cannot learn fine-grained correspondence relations. However, the existing methods still suffer from the noise contained in naive correlations and the lack of context semantic information in correlations. To alleviate these problems mentioned above, we propose a Feature-Enhanced Context-Aware Network (FECANet). Specifically, a feature enhancement module is proposed to suppress the matching noise caused by inter-class local similarity and enhance the intra-class relevance in the naive correlation. In addition, we propose a novel correlation reconstruction module that encodes extra correspondence relations between foreground and background and multi-scale context semantic features, significantly boosting the encoder to capture a reliable matching pattern. Experiments on PASCAL-$5^i$ and COCO-$20^i$ datasets demonstrate that our proposed FECANet leads to remarkable improvement compared to previous state-of-the-arts, demonstrating its effectiveness.
Neural Radiance Fields (NeRF) has achieved impressive results in single object scene reconstruction and novel view synthesis, which have been demonstrated on many single modality and single object focused indoor scene datasets like DTU, BMVS, and NeRF Synthetic.However, the study of NeRF on large-scale outdoor scene reconstruction is still limited, as there is no unified outdoor scene dataset for large-scale NeRF evaluation due to expensive data acquisition and calibration costs. In this paper, we propose a large-scale outdoor multi-modal dataset, OMMO dataset, containing complex land objects and scenes with calibrated images, point clouds and prompt annotations. Meanwhile, a new benchmark for several outdoor NeRF-based tasks is established, such as novel view synthesis, surface reconstruction, and multi-modal NeRF. To create the dataset, we capture and collect a large number of real fly-view videos and select high-quality and high-resolution clips from them. Then we design a quality review module to refine images, remove low-quality frames and fail-to-calibrate scenes through a learning-based automatic evaluation plus manual review. Finally, a number of volunteers are employed to add the text descriptions for each scene and key-frame to meet the potential multi-modal requirements in the future. Compared with existing NeRF datasets, our dataset contains abundant real-world urban and natural scenes with various scales, camera trajectories, and lighting conditions. Experiments show that our dataset can benchmark most state-of-the-art NeRF methods on different tasks. We will release the dataset and model weights very soon.
Neural Architecture Search has attracted increasing attention in recent years. Among them, differential NAS approaches such as DARTS, have gained popularity for the search efficiency. However, they still suffer from three main issues, that are, the weak stability due to the performance collapse, the poor generalization ability of the searched architectures, and the inferior robustness to different kinds of proxies. To solve the stability and generalization problems, a simple-but-effective regularization method, termed as Beta-Decay, is proposed to regularize the DARTS-based NAS searching process (i.e., $\beta$-DARTS). Specifically, Beta-Decay regularization can impose constraints to keep the value and variance of activated architecture parameters from being too large, thereby ensuring fair competition among architecture parameters and making the supernet less sensitive to the impact of input on the operation set. In-depth theoretical analyses on how it works and why it works are provided. Comprehensive experiments validate that Beta-Decay regularization can help to stabilize the searching process and makes the searched network more transferable across different datasets. To address the robustness problem, we first benchmark different NAS methods under a wide range of proxy data, proxy channels, proxy layers and proxy epochs, since the robustness of NAS under different kinds of proxies has not been explored before. We then conclude some interesting findings and find that $\beta$-DARTS always achieves the best result among all compared NAS methods under almost all proxies. We further introduce the novel flooding regularization to the weight optimization of $\beta$-DARTS (i.e., Bi-level regularization), and experimentally and theoretically verify its effectiveness for improving the proxy robustness of differentiable NAS.
Configurable software systems can be tuned for better performance. Leveraging on some Pareto optimizers, recent work has shifted from tuning for a single, time-related performance objective to two intrinsically different objectives that assess distinct performance aspects of the system, each with varying aspirations. Before we design better optimizers, a crucial engineering decision to make therein is how to handle the performance requirements with clear aspirations in the tuning process. For this, the community takes two alternative optimization models: either quantifying and incorporating the aspirations into the search objectives that guide the tuning, or not considering the aspirations during the search but purely using them in the later decision-making process only. However, despite being a crucial decision that determines how an optimizer can be designed and tailored, there is a rather limited understanding of which optimization model should be chosen under what particular circumstance, and why. In this paper, we seek to close this gap. Firstly, we do that through a review of over 426 papers in the literature and 14 real-world requirements datasets. Drawing on these, we then conduct a comprehensive empirical study that covers 15 combinations of the state-of-the-art performance requirement patterns, four types of aspiration space, three Pareto optimizers, and eight real-world systems/environments, leading to 1,296 cases of investigation. We found that (1) the realism of aspirations is the key factor that determines whether they should be used to guide the tuning; (2) the given patterns and the position of the realistic aspirations in the objective landscape are less important for the choice, but they do matter to the extents of improvement; (3) the available tuning budget can also influence the choice for unrealistic aspirations but it is insignificant under realistic ones.
3D dense captioning aims to generate multiple captions localized with their associated object regions. Existing methods follow a sophisticated ``detect-then-describe'' pipeline equipped with numerous hand-crafted components. However, these hand-crafted components would yield suboptimal performance given cluttered object spatial and class distributions among different scenes. In this paper, we propose a simple-yet-effective transformer framework Vote2Cap-DETR based on recent popular \textbf{DE}tection \textbf{TR}ansformer (DETR). Compared with prior arts, our framework has several appealing advantages: 1) Without resorting to numerous hand-crafted components, our method is based on a full transformer encoder-decoder architecture with a learnable vote query driven object decoder, and a caption decoder that produces the dense captions in a set-prediction manner. 2) In contrast to the two-stage scheme, our method can perform detection and captioning in one-stage. 3) Without bells and whistles, extensive experiments on two commonly used datasets, ScanRefer and Nr3D, demonstrate that our Vote2Cap-DETR surpasses current state-of-the-arts by 11.13\% and 7.11\% in CIDEr@0.5IoU, respectively. Codes will be released soon.
Medical Visual Question Answering (Medical-VQA) aims to to answer clinical questions regarding radiology images, assisting doctors with decision-making options. Nevertheless, current Medical-VQA models learn cross-modal representations through residing vision and texture encoders in dual separate spaces, which lead to indirect semantic alignment. In this paper, we propose UnICLAM, a Unified and Interpretable Medical-VQA model through Contrastive Representation Learning with Adversarial Masking. Specifically, to learn an aligned image-text representation, we first establish a unified dual-stream pre-training structure with the gradually soft-parameter sharing strategy. Technically, the proposed strategy learns a constraint for the vision and texture encoders to be close in a same space, which is gradually loosened as the higher number of layers. Moreover, for grasping the unified semantic representation, we extend the adversarial masking data augmentation to the contrastive representation learning of vision and text in a unified manner. Concretely, while the encoder training minimizes the distance between original and masking samples, the adversarial masking module keeps adversarial learning to conversely maximize the distance. Furthermore, we also intuitively take a further exploration to the unified adversarial masking augmentation model, which improves the potential ante-hoc interpretability with remarkable performance and efficiency. Experimental results on VQA-RAD and SLAKE public benchmarks demonstrate that UnICLAM outperforms existing 11 state-of-the-art Medical-VQA models. More importantly, we make an additional discussion about the performance of UnICLAM in diagnosing heart failure, verifying that UnICLAM exhibits superior few-shot adaption performance in practical disease diagnosis.
State-of-the-art 3D semantic segmentation models are trained on the off-the-shelf public benchmarks, but they often face the major challenge when these well-trained models are deployed to a new domain. In this paper, we propose an Active-and-Adaptive Segmentation (ADAS) baseline to enhance the weak cross-domain generalization ability of a well-trained 3D segmentation model, and bridge the point distribution gap between domains. Specifically, before the cross-domain adaptation stage begins, ADAS performs an active sampling operation to select a maximally-informative subset from both source and target domains for effective adaptation, reducing the adaptation difficulty under 3D scenarios. Benefiting from the rise of multi-modal 2D-3D datasets, ADAS utilizes a cross-modal attention-based feature fusion module that can extract a representative pair of image features and point features to achieve a bi-directional image-point feature interaction for better safe adaptation. Experimentally, ADAS is verified to be effective in many cross-domain settings including: 1) Unsupervised Domain Adaptation (UDA), which means that all samples from target domain are unlabeled; 2) Unsupervised Few-shot Domain Adaptation (UFDA) which means that only a few unlabeled samples are available in the unlabeled target domain; 3) Active Domain Adaptation (ADA) which means that the selected target samples by ADAS are manually annotated. Their results demonstrate that ADAS achieves a significant accuracy gain by easily coupling ADAS with self-training methods or off-the-shelf UDA works.
In this paper, we introduce a novel approach for ground plane normal estimation of wheeled vehicles. In practice, the ground plane is dynamically changed due to braking and unstable road surface. As a result, the vehicle pose, especially the pitch angle, is oscillating from subtle to obvious. Thus, estimating ground plane normal is meaningful since it can be encoded to improve the robustness of various autonomous driving tasks (e.g., 3D object detection, road surface reconstruction, and trajectory planning). Our proposed method only uses odometry as input and estimates accurate ground plane normal vectors in real time. Particularly, it fully utilizes the underlying connection between the ego pose odometry (ego-motion) and its nearby ground plane. Built on that, an Invariant Extended Kalman Filter (IEKF) is designed to estimate the normal vector in the sensor's coordinate. Thus, our proposed method is simple yet efficient and supports both camera- and inertial-based odometry algorithms. Its usability and the marked improvement of robustness are validated through multiple experiments on public datasets. For instance, we achieve state-of-the-art accuracy on KITTI dataset with the estimated vector error of 0.39{\deg}. Our code is available at github.com/manymuch/ground_normal_filter.
We study a challenging task, conditional human motion generation, which produces plausible human motion sequences according to various conditional inputs, such as action classes or textual descriptors. Since human motions are highly diverse and have a property of quite different distribution from conditional modalities, such as textual descriptors in natural languages, it is hard to learn a probabilistic mapping from the desired conditional modality to the human motion sequences. Besides, the raw motion data from the motion capture system might be redundant in sequences and contain noises; directly modeling the joint distribution over the raw motion sequences and conditional modalities would need a heavy computational overhead and might result in artifacts introduced by the captured noises. To learn a better representation of the various human motion sequences, we first design a powerful Variational AutoEncoder (VAE) and arrive at a representative and low-dimensional latent code for a human motion sequence. Then, instead of using a diffusion model to establish the connections between the raw motion sequences and the conditional inputs, we perform a diffusion process on the motion latent space. Our proposed Motion Latent-based Diffusion model (MLD) could produce vivid motion sequences conforming to the given conditional inputs and substantially reduce the computational overhead in both the training and inference stages. Extensive experiments on various human motion generation tasks demonstrate that our MLD achieves significant improvements over the state-of-the-art methods among extensive human motion generation tasks, with two orders of magnitude faster than previous diffusion models on raw motion sequences.