6D object pose estimation is a fundamental yet challenging problem in computer vision. Convolutional Neural Networks (CNNs) have recently proven to be capable of predicting reliable 6D pose estimates even under monocular settings. Nonetheless, CNNs are identified as being extremely data-driven, and acquiring adequate annotations is oftentimes very time-consuming and labor intensive. To overcome this limitation, we propose a novel monocular 6D pose estimation approach by means of self-supervised learning, removing the need for real annotations. After training our proposed network fully supervised with synthetic RGB data, we leverage current trends in noisy student training and differentiable rendering to further self-supervise the model on these unsupervised real RGB(-D) samples, seeking for a visually and geometrically optimal alignment. Moreover, employing both visible and amodal mask information, our self-supervision becomes very robust towards challenging scenarios such as occlusion. Extensive evaluations demonstrate that our proposed self-supervision outperforms all other methods relying on synthetic data or employing elaborate techniques from the domain adaptation realm. Noteworthy, our self-supervised approach consistently improves over its synthetically trained baseline and often almost closes the gap towards its fully supervised counterpart. The code and models are publicly available at https://github.com/THU-DA-6D-Pose-Group/self6dpp.git.
Transportation infrastructure, such as road or railroad networks, represent a fundamental component of our civilization. For sustainable planning and informed decision making, a thorough understanding of the long-term evolution of transportation infrastructure such as road networks is crucial. However, spatially explicit, multi-temporal road network data covering large spatial extents are scarce and rarely available prior to the 2000s. Herein, we propose a framework that employs increasingly available scanned and georeferenced historical map series to reconstruct past road networks, by integrating abundant, contemporary road network data and color information extracted from historical maps. Specifically, our method uses contemporary road segments as analytical units and extracts historical roads by inferring their existence in historical map series based on image processing and clustering techniques. We tested our method on over 300,000 road segments representing more than 50,000 km of the road network in the United States, extending across three study areas that cover 53 historical topographic map sheets dated between 1890 and 1950. We evaluated our approach by comparison to other historical datasets and against manually created reference data, achieving F-1 scores of up to 0.95, and showed that the extracted road network statistics are highly plausible over time, i.e., following general growth patterns. We demonstrated that contemporary geospatial data integrated with information extracted from historical map series open up new avenues for the quantitative analysis of long-term urbanization processes and landscape changes far beyond the era of operational remote sensing and digital cartography.
An intuitive way to search for images is to use queries composed of an example image and a complementary text. While the first provides rich and implicit context for the search, the latter explicitly calls for new traits, or specifies how some elements of the example image should be changed to retrieve the desired target image. Current approaches typically combine the features of each of the two elements of the query into a single representation, which can then be compared to the ones of the potential target images. Our work aims at shedding new light on the task by looking at it through the prism of two familiar and related frameworks: text-to-image and image-to-image retrieval. Taking inspiration from them, we exploit the specific relation of each query element with the targeted image and derive light-weight attention mechanisms which enable to mediate between the two complementary modalities. We validate our approach on several retrieval benchmarks, querying with images and their associated free-form text modifiers. Our method obtains state-of-the-art results without resorting to side information, multi-level features, heavy pre-training nor large architectures as in previous works.
Mutual information (MI) minimization has gained considerable interests in various machine learning tasks. However, estimating and minimizing MI in high-dimensional spaces remains a challenging problem, especially when only samples, rather than distribution forms, are accessible. Previous works mainly focus on MI lower bound approximation, which is not applicable to MI minimization problems. In this paper, we propose a novel Contrastive Log-ratio Upper Bound (CLUB) of mutual information. We provide a theoretical analysis of the properties of CLUB and its variational approximation. Based on this upper bound, we introduce an accelerated MI minimization training scheme, which bridges MI minimization with negative sampling. Simulation studies on Gaussian distributions show the reliable estimation ability of CLUB. Real-world MI minimization experiments, including domain adaptation and information bottleneck, further demonstrate the effectiveness of the proposed method.
Neural network based end-to-end Text-to-Speech (TTS) has greatly improved the quality of synthesized speech. While how to use massive spontaneous speech without transcription efficiently still remains an open problem. In this paper, we propose MHTTS, a fast multi-speaker TTS system that is robust to transcription errors and speaking style speech data. Specifically, we introduce a multi-head model and transfer text information from high-quality corpus with manual transcription to spontaneous speech with imperfectly recognized transcription by jointly training them. MHTTS has three advantages: 1) Our system synthesizes better quality multi-speaker voice with faster inference speed. 2) Our system is capable of transferring correct text information to data with imperfect transcription, simulated using corruption, or provided by an Automatic Speech Recogniser (ASR). 3) Our system can utilize massive real spontaneous speech with imperfect transcription and synthesize expressive voice.
Following the earlier formalism of the categorical representation learning (arXiv:2103.14770) by the first two authors, we discuss the construction of the "RG-flow based categorifier". Borrowing ideas from theory of renormalization group flows (RG) in quantum field theory, holographic duality, and hyperbolic geometry, and mixing them with neural ODE's, we construct a new algorithmic natural language processing (NLP) architecture, called the RG-flow categorifier or for short the RG categorifier, which is capable of data classification and generation in all layers. We apply our algorithmic platform to biomedical data sets and show its performance in the field of sequence-to-function mapping. In particular we apply the RG categorifier to particular genomic sequences of flu viruses and show how our technology is capable of extracting the information from given genomic sequences, find their hidden symmetries and dominant features, classify them and use the trained data to make stochastic prediction of new plausible generated sequences associated with new set of viruses which could avoid the human immune system. The content of the current article is part of the recent US patent application submitted by first two authors (U.S. Patent Application No.: 63/313.504).
For video frame interpolation (VFI), existing deep-learning-based approaches strongly rely on the ground-truth (GT) intermediate frames, which sometimes ignore the non-unique nature of motion judging from the given adjacent frames. As a result, these methods tend to produce averaged solutions that are not clear enough. To alleviate this issue, we propose to relax the requirement of reconstructing an intermediate frame as close to the GT as possible. Towards this end, we develop a texture consistency loss (TCL) upon the assumption that the interpolated content should maintain similar structures with their counterparts in the given frames. Predictions satisfying this constraint are encouraged, though they may differ from the pre-defined GT. Without the bells and whistles, our plug-and-play TCL is capable of improving the performance of existing VFI frameworks. On the other hand, previous methods usually adopt the cost volume or correlation map to achieve more accurate image/feature warping. However, the O(N^2) ({N refers to the pixel count}) computational complexity makes it infeasible for high-resolution cases. In this work, we design a simple, efficient (O(N)) yet powerful cross-scale pyramid alignment (CSPA) module, where multi-scale information is highly exploited. Extensive experiments justify the efficiency and effectiveness of the proposed strategy.
Graph Neural Networks (GNNs) show strong expressive power on graph data mining, by aggregating information from neighbors and using the integrated representation in the downstream tasks. The same aggregation methods and parameters for each node in a graph are used to enable the GNNs to utilize the homophily relational data. However, not all graphs are homophilic, even in the same graph, the distributions may vary significantly. Using the same convolution over all nodes may lead to the ignorance of various graph patterns. Furthermore, many existing GNNs integrate node features and structure identically, which ignores the distributions of nodes and further limits the expressive power of GNNs. To solve these problems, we propose Meta Weight Graph Neural Network (MWGNN) to adaptively construct graph convolution layers for different nodes. First, we model the Node Local Distribution (NLD) from node feature, topological structure and positional identity aspects with the Meta-Weight. Then, based on the Meta-Weight, we generate the adaptive graph convolutions to perform a node-specific weighted aggregation and boost the node representations. Finally, we design extensive experiments on real-world and synthetic benchmarks to evaluate the effectiveness of MWGNN. These experiments show the excellent expressive power of MWGNN in dealing with graph data with various distributions.
Most of the existing handcrafted and learning-based local descriptors are still at best approximately invariant to affine image transformations, often disregarding deformable surfaces. In this paper, we take one step further by proposing a new approach to compute descriptors from RGB-D images (where RGB refers to the pixel color brightness and D stands for depth information) that are invariant to isometric non-rigid deformations, as well as to scale changes and rotation. Our proposed description strategies are grounded on the key idea of learning feature representations on undistorted local image patches using surface geodesics. We design two complementary local descriptors strategies to compute geodesic-aware features efficiently: one efficient binary descriptor based on handcrafted binary tests (named GeoBit), and one learning-based descriptor (GeoPatch) with convolutional neural networks (CNNs) to compute features. In different experiments using real and publicly available RGB-D data benchmarks, they consistently outperforms state-of-the-art handcrafted and learning-based image and RGB-D descriptors in matching scores, as well as in object retrieval and non-rigid surface tracking experiments, with comparable processing times. We also provide to the community a new dataset with accurate matching annotations of RGB-D images of different objects (shirts, cloths, paintings, bags), subjected to strong non-rigid deformations, for evaluation benchmark of deformable surface correspondence algorithms.
Remote sensing change detection, identifying changes between scenes of the same location, is an active area of research with a broad range of applications. Recent advances in multimodal self-supervised pretraining have resulted in state-of-the-art methods which surpass vision models trained solely on optical imagery. In the remote sensing field, there is a wealth of overlapping 2D and 3D modalities which can be exploited to supervise representation learning in vision models. In this paper we propose Contrastive Surface-Image Pretraining (CSIP) for joint learning using optical RGB and above ground level (AGL) map pairs. We then evaluate these pretrained models on several building segmentation and change detection datasets to show that our method does, in fact, extract features relevant to downstream applications where natural and artificial surface information is relevant.