Point-Level Region Contrast for Object Detection Pre-Training

In this work we present point-level region contrast, a self-supervised pre-training approach for the task of object detection. This approach is motivated by the two key factors in detection: localization and recognition. While accurate localization favors models that operate at the pixel- or point-level, correct recognition typically relies on a more holistic, region-level view of objects. Incorporating this perspective in pre-training, our approach performs contrastive learning by directly sampling individual point pairs from different regions. Compared to an aggregated representation per region, our approach is more robust to the change in input region quality, and further enables us to implicitly improve initial region assignments via online knowledge distillation during training. Both advantages are important when dealing with imperfect regions encountered in the unsupervised setting. Experiments show point-level region contrast improves on state-of-the-art pre-training methods for object detection and segmentation across multiple tasks and datasets, and we provide extensive ablation studies and visualizations to aid understanding. Code will be made available.

Reducing Redundancy in the Bottleneck Representation of the Autoencoders

Autoencoders are a type of unsupervised neural networks, which can be used to solve various tasks, e.g., dimensionality reduction, image compression, and image denoising. An AE has two goals: (i) compress the original input to a low-dimensional space at the bottleneck of the network topology using an encoder, (ii) reconstruct the input from the representation at the bottleneck using a decoder. Both encoder and decoder are optimized jointly by minimizing a distortion-based loss which implicitly forces the model to keep only those variations of input data that are required to reconstruct the and to reduce redundancies. In this paper, we propose a scheme to explicitly penalize feature redundancies in the bottleneck representation. To this end, we propose an additional loss term, based on the pair-wise correlation of the neurons, which complements the standard reconstruction loss forcing the encoder to learn a more diverse and richer representation of the input. We tested our approach across different tasks: dimensionality reduction using three different dataset, image compression using the MNIST dataset, and image denoising using fashion MNIST. The experimental results show that the proposed loss leads consistently to superior performance compared to the standard AE loss.

Distance Estimation and Animal Tracking for Wildlife Camera Trapping

The ongoing biodiversity crysis calls for accurate estimation of animal density and abundance to identify, for example, sources of biodiversity decline and effectiveness of conservation interventions. Camera traps together with abundance estimation methods are often employed for this purpose. The necessary distances between camera and observed animal are traditionally derived in a laborious, fully manual or semi-automatic process. Both approaches require reference image material, which is both difficult to acquire and not available for existing datasets. In this study, we propose a fully automatic approach to estimate camera-to-animal distances, based on monocular depth estimation (MDE), and without the need of reference image material. We leverage state-of-the-art relative MDE and a novel alignment procedure to estimate metric distances. We evaluate the approach on a zoo scenario dataset unseen during training. We achieve a mean absolute distance estimation error of only 0.9864 meters at a precision of 90.3% and recall of 63.8%, while completely eliminating the previously required manual effort for biodiversity researchers. The code will be made available.

NIMBLE: A Non-rigid Hand Model with Bones and Muscles

Emerging Metaverse applications demand reliable, accurate, and photorealistic reproductions of human hands to perform sophisticated operations as if in the physical world. While real human hand represents one of the most intricate coordination between bones, muscle, tendon, and skin, state-of-the-art techniques unanimously focus on modeling only the skeleton of the hand. In this paper, we present NIMBLE, a novel parametric hand model that includes the missing key components, bringing 3D hand model to a new level of realism. We first annotate muscles, bones and skins on the recent Magnetic Resonance Imaging hand (MRI-Hand) dataset and then register a volumetric template hand onto individual poses and subjects within the dataset. NIMBLE consists of 20 bones as triangular meshes, 7 muscle groups as tetrahedral meshes, and a skin mesh. Via iterative shape registration and parameter learning, it further produces shape blend shapes, pose blend shapes, and a joint regressor. We demonstrate applying NIMBLE to modeling, rendering, and visual inference tasks. By enforcing the inner bones and muscles to match anatomic and kinematic rules, NIMBLE can animate 3D hands to new poses at unprecedented realism. To model the appearance of skin, we further construct a photometric HandStage to acquire high-quality textures and normal maps to model wrinkles and palm print. Finally, NIMBLE also benefits learning-based hand pose and shape estimation by either synthesizing rich data or acting directly as a differentiable layer in the inference network.

CRAT-Pred: Vehicle Trajectory Prediction with Crystal Graph Convolutional Neural Networks and Multi-Head Self-Attention

Predicting the motion of surrounding vehicles is essential for autonomous vehicles, as it governs their own motion plan. Current state-of-the-art vehicle prediction models heavily rely on map information. In reality, however, this information is not always available. We therefore propose CRAT-Pred, a multi-modal and non-rasterization-based trajectory prediction model, specifically designed to effectively model social interactions between vehicles, without relying on map information. CRAT-Pred applies a graph convolution method originating from the field of material science to vehicle prediction, allowing to efficiently leverage edge features, and combines it with multi-head self-attention. Compared to other map-free approaches, the model achieves state-of-the-art performance with a significantly lower number of model parameters. In addition to that, we quantitatively show that the self-attention mechanism is able to learn social interactions between vehicles, with the weights representing a measurable interaction score. The source code is publicly available.

Object-Guided Day-Night Visual Localization in Urban Scenes

We introduce Object-Guided Localization (OGuL) based on a novel method of local-feature matching. Direct matching of local features is sensitive to significant changes in illumination. In contrast, object detection often survives severe changes in lighting conditions. The proposed method first detects semantic objects and establishes correspondences of those objects between images. Object correspondences provide local coarse alignment of the images in the form of a planar homography. These homographies are consequently used to guide the matching of local features. Experiments on standard urban localization datasets (Aachen, Extended-CMU-Season, RobotCar-Season) show that OGuL significantly improves localization results with as simple local features as SIFT, and its performance competes with the state-of-the-art CNN-based methods trained for day-to-night localization.

Deep Feature Rotation for Multimodal Image Style Transfer

Recently, style transfer is a research area that attracts a lot of attention, which transfers the style of an image onto a content target. Extensive research on style transfer has aimed at speeding up processing or generating high-quality stylized images. Most approaches only produce an output from a content and style image pair, while a few others use complex architectures and can only produce a certain number of outputs. In this paper, we propose a simple method for representing style features in many ways called Deep Feature Rotation (DFR), while not only producing diverse outputs but also still achieving effective stylization compared to more complex methods. Our approach is representative of the many ways of augmentation for intermediate feature embedding without consuming too much computational expense. We also analyze our method by visualizing output in different rotation weights. Our code is available at

Anchor Graph Structure Fusion Hashing for Cross-Modal Similarity Search

Cross-modal hashing still has some challenges needed to address: (1) most existing CMH methods take graphs as input to model data distribution. These methods omit to consider the correlation of graph structure among multiple modalities; (2) most existing CMH methods ignores considering the fusion affinity among multi-modalities data; (3) most existing CMH methods relax the discrete constraints to solve the optimization objective, significantly degrading the retrieval performance. To solve the above limitations, we propose a novel Anchor Graph Structure Fusion Hashing (AGSFH). AGSFH constructs the anchor graph structure fusion matrix from different anchor graphs of multiple modalities with the Hadamard product, which can fully exploit the geometric property of underlying data structure. Based on the anchor graph structure fusion matrix, AGSFH attempts to directly learn an intrinsic anchor graph, where the structure of the intrinsic anchor graph is adaptively tuned so that the number of components of the intrinsic graph is exactly equal to the number of clusters. Besides, AGSFH preserves the anchor fusion affinity into the common binary Hamming space. Furthermore, a discrete optimization framework is designed to learn the unified binary codes. Extensive experimental results on three public social datasets demonstrate the superiority of AGSFH.

Conditional Motion In-betweening

Motion in-betweening (MIB) is a process of generating intermediate skeletal movement between the given start and target poses while preserving the naturalness of the motion, such as periodic footstep motion while walking. Although state-of-the-art MIB methods are capable of producing plausible motions given sparse key-poses, they often lack the controllability to generate motions satisfying the semantic contexts required in practical applications. We focus on the method that can handle pose or semantic conditioned MIB tasks using a unified model. We also present a motion augmentation method to improve the quality of pose-conditioned motion generation via defining a distribution over smooth trajectories. Our proposed method outperforms the existing state-of-the-art MIB method in pose prediction errors while providing additional controllability.

Learning to Bootstrap for Combating Label Noise

Deep neural networks are powerful tools for representation learning, but can easily overfit to noisy labels which are prevalent in many real-world scenarios. Generally, noisy supervision could stem from variation among labelers, label corruption by adversaries, etc. To combat such label noises, one popular line of approach is to apply customized weights to the training instances, so that the corrupted examples contribute less to the model learning. However, such learning mechanisms potentially erase important information about the data distribution and therefore yield suboptimal results. To leverage useful information from the corrupted instances, an alternative is the bootstrapping loss, which reconstructs new training targets on-the-fly by incorporating the network’s own predictions (i.e., pseudo-labels).