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.
Despite advancements in user-guided video segmentation, extracting complex objects consistently for highly complex scenes is still a labor-intensive task, especially for production. It is not uncommon that a majority of frames need to be annotated. We introduce a novel semi-supervised video object segmentation (SSVOS) model, XMem++, that improves existing memory-based models, with a permanent memory module. Most existing methods focus on single frame annotations, while our approach can effectively handle multiple user-selected frames with varying appearances of the same object or region. Our method can extract highly consistent results while keeping the required number of frame annotations low. We further introduce an iterative and attention-based frame suggestion mechanism, which computes the next best frame for annotation. Our method is real-time and does not require retraining after each user input. We also introduce a new dataset, PUMaVOS, which covers new challenging use cases not found in previous benchmarks. We demonstrate SOTA performance on challenging (partial and multi-class) segmentation scenarios as well as long videos, while ensuring significantly fewer frame annotations than any existing method. Project page: https://max810.github.io/xmem2-project-page/
Humans have the ability to reuse previously learned policies to solve new tasks quickly, and reinforcement learning (RL) agents can do the same by transferring knowledge from source policies to a related target task. Transfer RL methods can reshape the policy optimization objective (optimization transfer) or influence the behavior policy (behavior transfer) using source policies. However, selecting the appropriate source policy with limited samples to guide target policy learning has been a challenge. Previous methods introduce additional components, such as hierarchical policies or estimations of source policies' value functions, which can lead to non-stationary policy optimization or heavy sampling costs, diminishing transfer effectiveness. To address this challenge, we propose a novel transfer RL method that selects the source policy without training extra components. Our method utilizes the Q function in the actor-critic framework to guide policy selection, choosing the source policy with the largest one-step improvement over the current target policy. We integrate optimization transfer and behavior transfer (IOB) by regularizing the learned policy to mimic the guidance policy and combining them as the behavior policy. This integration significantly enhances transfer effectiveness, surpasses state-of-the-art transfer RL baselines in benchmark tasks, and improves final performance and knowledge transferability in continual learning scenarios. Additionally, we show that our optimization transfer technique is guaranteed to improve target policy learning.
Convolutional Neural Networks (CNNs) have exhibited strong performance in medical image segmentation tasks by capturing high-level (local) information, such as edges and textures. However, due to the limited field of view of convolution kernel, it is hard for CNNs to fully represent global information. Recently, transformers have shown good performance for medical image segmentation due to their ability to better model long-range dependencies. Nevertheless, transformers struggle to capture high-level spatial features as effectively as CNNs. A good segmentation model should learn a better representation from local and global features to be both precise and semantically accurate. In our previous work, we proposed CATS, which is a U-shaped segmentation network augmented with transformer encoder. In this work, we further extend this model and propose CATS v2 with hybrid encoders. Specifically, hybrid encoders consist of a CNN-based encoder path paralleled to a transformer path with a shifted window, which better leverage both local and global information to produce robust 3D medical image segmentation. We fuse the information from the convolutional encoder and the transformer at the skip connections of different resolutions to form the final segmentation. The proposed method is evaluated on two public challenge datasets: Cross-Modality Domain Adaptation (CrossMoDA) and task 5 of Medical Segmentation Decathlon (MSD-5), to segment vestibular schwannoma (VS) and prostate, respectively. Compared with the state-of-the-art methods, our approach demonstrates superior performance in terms of higher Dice scores.
We present the All-Seeing (AS) project: a large-scale data and model for recognizing and understanding everything in the open world. Using a scalable data engine that incorporates human feedback and efficient models in the loop, we create a new dataset (AS-1B) with over 1 billion regions annotated with semantic tags, question-answering pairs, and detailed captions. It covers a wide range of 3.5 million common and rare concepts in the real world, and has 132.2 billion tokens that describe the concepts and their attributes. Leveraging this new dataset, we develop the All-Seeing model (ASM), a unified framework for panoptic visual recognition and understanding. The model is trained with open-ended language prompts and locations, which allows it to generalize to various vision and language tasks with remarkable zero-shot performance, including region-text retrieval, region recognition, captioning, and question-answering. We hope that this project can serve as a foundation for vision-language artificial general intelligence research. Models and the dataset shall be released at https://github.com/OpenGVLab/All-Seeing, and demo can be seen at https://huggingface.co/spaces/OpenGVLab/all-seeing.
The Traveling Salesman Problem (TSP) is a well-known problem in combinatorial optimization with applications in various domains. However, existing TSP solvers face challenges in producing high-quality solutions with low latency. To address this issue, we propose NAR4TSP, which produces TSP solutions in a Non-Autoregressive (NAR) manner using a specially designed Graph Neural Network (GNN), achieving faster inference speed. Moreover, NAR4TSP is trained using an enhanced Reinforcement Learning (RL) strategy, eliminating the dependency on costly labels used to train conventional supervised learning-based NAR models. To the best of our knowledge, NAR4TSP is the first TSP solver that successfully combines RL and NAR decoding. The experimental results on both synthetic and real-world TSP instances demonstrate that NAR4TSP outperforms four state-of-the-art models in terms of solution quality, inference latency, and generalization ability. Lastly, we present visualizations of NAR4TSP's decoding process and its overall path planning to showcase the feasibility of implementing NAR4TSP in an end-to-end manner and its effectiveness, respectively.
Recently, there has been a growing interest in text-to-speech (TTS) methods that can be trained with minimal supervision by combining two types of discrete speech representations and using two sequence-to-sequence tasks to decouple TTS. To address the challenges associated with high dimensionality and waveform distortion in discrete representations, we propose Diff-LM-Speech, which models semantic embeddings into mel-spectrogram based on diffusion models and introduces a prompt encoder structure based on variational autoencoders and prosody bottlenecks to improve prompt representation capabilities. Autoregressive language models often suffer from missing and repeated words, while non-autoregressive frameworks face expression averaging problems due to duration prediction models. To address these issues, we propose Tetra-Diff-Speech, which designs a duration diffusion model to achieve diverse prosodic expressions. While we expect the information content of semantic coding to be between that of text and acoustic coding, existing models extract semantic coding with a lot of redundant information and dimensionality explosion. To verify that semantic coding is not necessary, we propose Tri-Diff-Speech. Experimental results show that our proposed methods outperform baseline methods. We provide a website with audio samples.
Medical image segmentation is a critical task in medical image analysis. In recent years, deep learning based approaches have shown exceptional performance when trained on a fully-annotated dataset. However, data annotation is often a significant bottleneck, especially for 3D medical images. Active learning (AL) is a promising solution for efficient annotation but requires an initial set of labeled samples to start active selection. When the entire data pool is unlabeled, how do we select the samples to annotate as our initial set? This is also known as the cold-start AL, which permits only one chance to request annotations from experts without access to previously annotated data. Cold-start AL is highly relevant in many practical scenarios but has been under-explored, especially for 3D medical segmentation tasks requiring substantial annotation effort. In this paper, we present a benchmark named COLosSAL by evaluating six cold-start AL strategies on five 3D medical image segmentation tasks from the public Medical Segmentation Decathlon collection. We perform a thorough performance analysis and explore important open questions for cold-start AL, such as the impact of budget on different strategies. Our results show that cold-start AL is still an unsolved problem for 3D segmentation tasks but some important trends have been observed. The code repository, data partitions, and baseline results for the complete benchmark are publicly available at https://github.com/MedICL-VU/COLosSAL.
Accurate building height estimation is key to the automatic derivation of 3D city models from emerging big geospatial data, including Volunteered Geographical Information (VGI). However, an automatic solution for large-scale building height estimation based on low-cost VGI data is currently missing. The fast development of VGI data platforms, especially OpenStreetMap (OSM) and crowdsourced street-view images (SVI), offers a stimulating opportunity to fill this research gap. In this work, we propose a semi-supervised learning (SSL) method of automatically estimating building height from Mapillary SVI and OSM data to generate low-cost and open-source 3D city modeling in LoD1. The proposed method consists of three parts: first, we propose an SSL schema with the option of setting a different ratio of "pseudo label" during the supervised regression; second, we extract multi-level morphometric features from OSM data (i.e., buildings and streets) for the purposed of inferring building height; last, we design a building floor estimation workflow with a pre-trained facade object detection network to generate "pseudo label" from SVI and assign it to the corresponding OSM building footprint. In a case study, we validate the proposed SSL method in the city of Heidelberg, Germany and evaluate the model performance against the reference data of building heights. Based on three different regression models, namely Random Forest (RF), Support Vector Machine (SVM), and Convolutional Neural Network (CNN), the SSL method leads to a clear performance boosting in estimating building heights with a Mean Absolute Error (MAE) around 2.1 meters, which is competitive to state-of-the-art approaches. The preliminary result is promising and motivates our future work in scaling up the proposed method based on low-cost VGI data, with possibilities in even regions and areas with diverse data quality and availability.