Computer Vision and Pattern Recognition (cs.CV)

Abstract: Unmanned Aerial Vehicles (UAVs), specifically drones equipped with remote sensing object detection technology, have rapidly gained a broad spectrum of applications and emerged as one of the primary research focuses in the field of computer vision. Although UAV remote sensing systems have the ability to detect various objects, small-scale objects can be challenging to detect reliably due to factors such as object size, image degradation, and real-time limitations. To tackle these issues, a real-time object detection algorithm (YOLO-Drone) is proposed and applied to two new UAV platforms as well as a specific light source (silicon-based golden LED). YOLO-Drone presents several novelties: 1) including a new backbone Darknet59; 2) a new complex feature aggregation module MSPP-FPN that incorporated one spatial pyramid pooling and three atrous spatial pyramid pooling modules; 3) and the use of Generalized Intersection over Union (GIoU) as the loss function. To evaluate performance, two benchmark datasets, UAVDT and VisDrone, along with one homemade dataset acquired at night under silicon-based golden LEDs, are utilized. The experimental results show that, in both UAVDT and VisDrone, the proposed YOLO-Drone outperforms state-of-the-art (SOTA) object detection methods by improving the mAP of 10.13% and 8.59%, respectively. With regards to UAVDT, the YOLO-Drone exhibits both high real-time inference speed of 53 FPS and a maximum mAP of 34.04%. Notably, YOLO-Drone achieves high performance under the silicon-based golden LEDs, with a mAP of up to 87.71%, surpassing the performance of YOLO series under ordinary light sources. To conclude, the proposed YOLO-Drone is a highly effective solution for object detection in UAV applications, particularly for night detection tasks where silicon-based golden light LED technology exhibits significant superiority.

Abstract: We tackle the issue of generalized category discovery (GCD). GCD considers the open-world problem of automatically clustering a partially labelled dataset, in which the unlabelled data contain instances from novel categories and also the labelled classes. In this paper, we address the GCD problem without a known category number in the unlabelled data. We propose a framework, named CiPR, to bootstrap the representation by exploiting Cross-instance Positive Relations for contrastive learning in the partially labelled data which are neglected in existing methods. First, to obtain reliable cross-instance relations to facilitate the representation learning, we introduce a semi-supervised hierarchical clustering algorithm, named selective neighbor clustering (SNC), which can produce a clustering hierarchy directly from the connected components in the graph constructed by selective neighbors. We also extend SNC to be capable of label assignment for the unlabelled instances with the given class number. Moreover, we present a method to estimate the unknown class number using SNC with a joint reference score considering clustering indexes of both labelled and unlabelled data. Finally, we thoroughly evaluate our framework on public generic image recognition datasets and challenging fine-grained datasets, all establishing the new state-of-the-art.

Abstract: Scene Dynamic Recovery (SDR) by inverting distorted Rolling Shutter (RS) images to an undistorted high frame-rate Global Shutter (GS) video is a severely ill-posed problem, particularly when prior knowledge about camera/object motions is unavailable. Commonly used artificial assumptions on motion linearity and data-specific characteristics, regarding the temporal dynamics information embedded in the RS scanlines, are prone to producing sub-optimal solutions in real-world scenarios. To address this challenge, we propose an event-based RS2GS framework within a self-supervised learning paradigm that leverages the extremely high temporal resolution of event cameras to provide accurate inter/intra-frame information. % In this paper, we propose to leverage the event camera to provide inter/intra-frame information as the emitted events have an extremely high temporal resolution and learn an event-based RS2GS network within a self-supervised learning framework, where real-world events and RS images can be exploited to alleviate the performance degradation caused by the domain gap between the synthesized and real data. Specifically, an Event-based Inter/intra-frame Compensator (E-IC) is proposed to predict the per-pixel dynamic between arbitrary time intervals, including the temporal transition and spatial translation. Exploring connections in terms of RS-RS, RS-GS, and GS-RS, we explicitly formulate mutual constraints with the proposed E-IC, resulting in supervisions without ground-truth GS images. Extensive evaluations over synthetic and real datasets demonstrate that the proposed method achieves state-of-the-art and shows remarkable performance for event-based RS2GS inversion in real-world scenarios. The dataset and code are available at https://w3un.github.io/selfunroll/.

Abstract: Federated learning (FL) has been introduced to the healthcare domain as a decentralized learning paradigm that allows multiple parties to train a model collaboratively without privacy leakage. However, most previous studies have assumed that every client holds an identical label set. In reality, medical specialists tend to annotate only diseases within their knowledge domain or interest. This implies that label sets in each client can be different and even disjoint. In this paper, we propose the framework FedLSM to solve the problem Label Set Mismatch. FedLSM adopts different training strategies on data with different uncertainty levels to efficiently utilize unlabeled or partially labeled data as well as class-wise adaptive aggregation in the classification layer to avoid inaccurate aggregation when clients have missing labels. We evaluate FedLSM on two public real-world medical image datasets, including chest x-ray (CXR) diagnosis with 112,120 CXR images and skin lesion diagnosis with 10,015 dermoscopy images, and show that it significantly outperforms other state-of-the-art FL algorithms. Code will be made available upon acceptance.

Abstract: Animating an object in 3D often requires an articulated structure, e.g. a kinematic chain or skeleton of the manipulated object with proper skinning weights, to obtain smooth movements and surface deformations. However, existing models that allow direct pose manipulations are either limited to specific object categories or built with specialized equipment. To reduce the work needed for creating animatable 3D models, we propose a novel reconstruction method that learns an animatable kinematic chain for any articulated object. Our method operates on monocular videos without prior knowledge of the object's shape or underlying structure. Our approach is on par with state-of-the-art 3D surface reconstruction methods on various articulated object categories while enabling direct pose manipulations by re-posing the learned kinematic chain.

Abstract: In-context vision and language models like Flamingo support arbitrarily interleaved sequences of images and text as input. This format not only enables few-shot learning via interleaving independent supervised (image, text) examples, but also, more complex prompts involving interaction between images, e.g., "What do image A and image B have in common?" To support this interface, pretraining occurs over web corpora that similarly contain interleaved images+text. To date, however, large-scale data of this form have not been publicly available. We release Multimodal C4 (mmc4), an augmentation of the popular text-only c4 corpus with images interleaved. We use a linear assignment algorithm to place images into longer bodies of text using CLIP features, a process that we show outperforms alternatives. mmc4 spans everyday topics like cooking, travel, technology, etc. A manual inspection of a random sample of documents shows that a vast majority (90%) of images are topically relevant, and that linear assignment frequently selects individual sentences specifically well-aligned with each image (78%). After filtering NSFW images, ads, etc., the corpus contains 103M documents containing 585M images interleaved with 43B English tokens.

Abstract: Mapping Low Dynamic Range (LDR) images with different exposures to High Dynamic Range (HDR) remains nontrivial and challenging on dynamic scenes due to ghosting caused by object motion or camera jitting. With the success of Deep Neural Networks (DNNs), several DNNs-based methods have been proposed to alleviate ghosting, they cannot generate approving results when motion and saturation occur. To generate visually pleasing HDR images in various cases, we propose a hybrid HDR deghosting network, called HyHDRNet, to learn the complicated relationship between reference and non-reference images. The proposed HyHDRNet consists of a content alignment subnetwork and a Transformer-based fusion subnetwork. Specifically, to effectively avoid ghosting from the source, the content alignment subnetwork uses patch aggregation and ghost attention to integrate similar content from other non-reference images with patch level and suppress undesired components with pixel level. To achieve mutual guidance between patch-level and pixel-level, we leverage a gating module to sufficiently swap useful information both in ghosted and saturated regions. Furthermore, to obtain a high-quality HDR image, the Transformer-based fusion subnetwork uses a Residual Deformable Transformer Block (RDTB) to adaptively merge information for different exposed regions. We examined the proposed method on four widely used public HDR image deghosting datasets. Experiments demonstrate that HyHDRNet outperforms state-of-the-art methods both quantitatively and qualitatively, achieving appealing HDR visualization with unified textures and colors.

Abstract: Reconstructing an image from noisy and incomplete measurements is a central task in several image processing applications. In recent years, state-of-the-art reconstruction methods have been developed based on recent advances in deep learning. Especially for highly underdetermined problems, maintaining data consistency is a key goal. This can be achieved either by iterative network architectures or by a subsequent projection of the network reconstruction. However, for such approaches to be used in safety-critical domains such as medical imaging, the network reconstruction should not only provide the user with a reconstructed image, but also with some level of confidence in the reconstruction. In order to meet these two key requirements, this paper combines deep null-space networks with uncertainty quantification. Evaluation of the proposed method includes image reconstruction from undersampled Radon measurements on a toy CT dataset and accelerated MRI reconstruction on the fastMRI dataset. This work is the first approach to solving inverse problems that additionally models data-dependent uncertainty by estimating an input-dependent scale map, providing a robust assessment of reconstruction quality.

Abstract: CLIP (Contrastive Language-Image Pretraining) is well-developed for open-vocabulary zero-shot image-level recognition, while its applications in pixel-level tasks are less investigated, where most efforts directly adopt CLIP features without deliberative adaptations. In this work, we first demonstrate the necessity of image-pixel CLIP feature adaption, then provide Multi-View Prompt learning (MVP-SEG) as an effective solution to achieve image-pixel adaptation and to solve open-vocabulary semantic segmentation. Concretely, MVP-SEG deliberately learns multiple prompts trained by our Orthogonal Constraint Loss (OCLoss), by which each prompt is supervised to exploit CLIP feature on different object parts, and collaborative segmentation masks generated by all prompts promote better segmentation. Moreover, MVP-SEG introduces Global Prompt Refining (GPR) to further eliminate class-wise segmentation noise. Experiments show that the multi-view prompts learned from seen categories have strong generalization to unseen categories, and MVP-SEG+ which combines the knowledge transfer stage significantly outperforms previous methods on several benchmarks. Moreover, qualitative results justify that MVP-SEG does lead to better focus on different local parts.

Abstract: Depth Estimation has wide reaching applications in the field of Computer vision such as target tracking, augmented reality, and self-driving cars. The goal of Monocular Depth Estimation is to predict the depth map, given a 2D monocular RGB image as input. The traditional depth estimation methods are based on depth cues and used concepts like epipolar geometry. With the evolution of Convolutional Neural Networks, depth estimation has undergone tremendous strides. In this project, our aim is to explore possible extensions to existing SoTA Deep Learning based Depth Estimation Models and to see whether performance metrics could be further improved. In a broader sense, we are looking at the possibility of implementing Pose Estimation, Efficient Sub-Pixel Convolution Interpolation, Semantic Segmentation Estimation techniques to further enhance our proposed architecture and to provide fine-grained and more globally coherent depth map predictions. We also plan to do away with camera intrinsic parameters during training and apply weather augmentations to further generalize our model.

Abstract: The limited ability of Convolutional Neural Networks to generalize to images from previously unseen domains is a major limitation, in particular, for safety-critical clinical tasks such as dermoscopic skin cancer classification. In order to translate CNN-based applications into the clinic, it is essential that they are able to adapt to domain shifts. Such new conditions can arise through the use of different image acquisition systems or varying lighting conditions. In dermoscopy, shifts can also occur as a change in patient age or occurence of rare lesion localizations (e.g. palms). These are not prominently represented in most training datasets and can therefore lead to a decrease in performance. In order to verify the generalizability of classification models in real world clinical settings it is crucial to have access to data which mimics such domain shifts. To our knowledge no dermoscopic image dataset exists where such domain shifts are properly described and quantified. We therefore grouped publicly available images from ISIC archive based on their metadata (e.g. acquisition location, lesion localization, patient age) to generate meaningful domains. To verify that these domains are in fact distinct, we used multiple quantification measures to estimate the presence and intensity of domain shifts. Additionally, we analyzed the performance on these domains with and without an unsupervised domain adaptation technique. We observed that in most of our grouped domains, domain shifts in fact exist. Based on our results, we believe these datasets to be helpful for testing the generalization capabilities of dermoscopic skin cancer classifiers.

Abstract: Neural radiance field (NeRF) has become a popular 3D representation method for human avatar reconstruction due to its high-quality rendering capabilities, e.g., regarding novel views and poses. However, previous methods for editing the geometry and appearance of the avatar only allow for global editing through body shape parameters and 2D texture maps. In this paper, we propose a new approach named \textbf{U}nified \textbf{V}olumetric \textbf{A}vatar (\textbf{UVA}) that enables local and independent editing of both geometry and texture, while retaining the ability to render novel views and poses. UVA transforms each observation point to a canonical space using a skinning motion field and represents geometry and texture in separate neural fields. Each field is composed of a set of structured latent codes that are attached to anchor nodes on a deformable mesh in canonical space and diffused into the entire space via interpolation, allowing for local editing. To address spatial ambiguity in code interpolation, we use a local signed height indicator. We also replace the view-dependent radiance color with a pose-dependent shading factor to better represent surface illumination in different poses. Experiments on multiple human avatars demonstrate that our UVA achieves competitive results in novel view synthesis and novel pose rendering while enabling local and independent editing of geometry and appearance. The source code will be released.

Abstract: In this paper, we realize automatic visual recognition and direction estimation of pointing. We introduce the first neural pointing understanding method based on two key contributions. The first is the introduction of a first-of-its-kind large-scale dataset for pointing recognition and direction estimation, which we refer to as the DP Dataset. DP Dataset consists of more than 2 million frames of over 33 people pointing in various styles annotated for each frame with pointing timings and 3D directions. The second is DeePoint, a novel deep network model for joint recognition and 3D direction estimation of pointing. DeePoint is a Transformer-based network which fully leverages the spatio-temporal coordination of the body parts, not just the hands. Through extensive experiments, we demonstrate the accuracy and efficiency of DeePoint. We believe DP Dataset and DeePoint will serve as a sound foundation for visual human intention understanding.

Abstract: File fragment classification (FFC) on small chunks of memory is essential in memory forensics and Internet security. Existing methods mainly treat file fragments as 1d byte signals and utilize the captured inter-byte features for classification, while the bit information within bytes, i.e., intra-byte information, is seldom considered. This is inherently inapt for classifying variable-length coding files whose symbols are represented as the variable number of bits. Conversely, we propose Byte2Image, a novel data augmentation technique, to introduce the neglected intra-byte information into file fragments and re-treat them as 2d gray-scale images, which allows us to capture both inter-byte and intra-byte correlations simultaneously through powerful convolutional neural networks (CNNs). Specifically, to convert file fragments to 2d images, we employ a sliding byte window to expose the neglected intra-byte information and stack their n-gram features row by row. We further propose a byte sequence \& image fusion network as a classifier, which can jointly model the raw 1d byte sequence and the converted 2d image to perform FFC. Experiments on FFT-75 dataset validate that our proposed method can achieve notable accuracy improvements over state-of-the-art methods in nearly all scenarios. The code will be released at https://github.com/wenyang001/Byte2Image.

Abstract: Efficient deep learning-based approaches have achieved remarkable performance in single image super-resolution. However, recent studies on efficient super-resolution have mainly focused on reducing the number of parameters and floating-point operations through various network designs. Although these methods can decrease the number of parameters and floating-point operations, they may not necessarily reduce actual running time. To address this issue, we propose a novel multi-stage lightweight network boosting method, which can enable lightweight networks to achieve outstanding performance. Specifically, we leverage enhanced high-resolution output as additional supervision to improve the learning ability of lightweight student networks. Upon convergence of the student network, we further simplify our network structure to a more lightweight level using reparameterization techniques and iterative network pruning. Meanwhile, we adopt an effective lightweight network training strategy that combines multi-anchor distillation and progressive learning, enabling the lightweight network to achieve outstanding performance. Ultimately, our proposed method achieves the fastest inference time among all participants in the NTIRE 2023 efficient super-resolution challenge while maintaining competitive super-resolution performance. Additionally, extensive experiments are conducted to demonstrate the effectiveness of the proposed components. The results show that our approach achieves comparable performance in representative dataset DIV2K, both qualitatively and quantitatively, with faster inference and fewer number of network parameters.

Abstract: The exploitation of visible spectrum datasets has led deep networks to show remarkable success. However, real-world tasks include low-lighting conditions which arise performance bottlenecks for models trained on large-scale RGB image datasets. Thermal IR cameras are more robust against such conditions. Therefore, the usage of thermal imagery in real-world applications can be useful. Unsupervised domain adaptation (UDA) allows transferring information from a source domain to a fully unlabeled target domain. Despite substantial improvements in UDA, the performance gap between UDA and its supervised learning counterpart remains significant. By picking a small number of target samples to annotate and using them in training, active domain adaptation tries to mitigate this gap with minimum annotation expense. We propose an active domain adaptation method in order to examine the efficiency of combining the visible spectrum and thermal imagery modalities. When the domain gap is considerably large as in the visible-to-thermal task, we may conclude that the methods without explicit domain alignment cannot achieve their full potential. To this end, we propose a spectral transfer guided active domain adaptation method to select the most informative unlabeled target samples while aligning source and target domains. We used the large-scale visible spectrum dataset MS-COCO as the source domain and the thermal dataset FLIR ADAS as the target domain to present the results of our method. Extensive experimental evaluation demonstrates that our proposed method outperforms the state-of-the-art active domain adaptation methods. The code and models are publicly available.

Abstract: Low-light image enhancement (LLIE) investigates how to improve illumination and produce normal-light images. The majority of existing methods improve low-light images via a global and uniform manner, without taking into account the semantic information of different regions. Without semantic priors, a network may easily deviate from a region's original color. To address this issue, we propose a novel semantic-aware knowledge-guided framework (SKF) that can assist a low-light enhancement model in learning rich and diverse priors encapsulated in a semantic segmentation model. We concentrate on incorporating semantic knowledge from three key aspects: a semantic-aware embedding module that wisely integrates semantic priors in feature representation space, a semantic-guided color histogram loss that preserves color consistency of various instances, and a semantic-guided adversarial loss that produces more natural textures by semantic priors. Our SKF is appealing in acting as a general framework in LLIE task. Extensive experiments show that models equipped with the SKF significantly outperform the baselines on multiple datasets and our SKF generalizes to different models and scenes well. The code is available at Semantic-Aware-Low-Light-Image-Enhancement.

Abstract: This paper discusses the results for the second edition of the Monocular Depth Estimation Challenge (MDEC). This edition was open to methods using any form of supervision, including fully-supervised, self-supervised, multi-task or proxy depth. The challenge was based around the SYNS-Patches dataset, which features a wide diversity of environments with high-quality dense ground-truth. This includes complex natural environments, e.g. forests or fields, which are greatly underrepresented in current benchmarks. The challenge received eight unique submissions that outperformed the provided SotA baseline on any of the pointcloud- or image-based metrics. The top supervised submission improved relative F-Score by 27.62%, while the top self-supervised improved it by 16.61%. Supervised submissions generally leveraged large collections of datasets to improve data diversity. Self-supervised submissions instead updated the network architecture and pretrained backbones. These results represent a significant progress in the field, while highlighting avenues for future research, such as reducing interpolation artifacts at depth boundaries, improving self-supervised indoor performance and overall natural image accuracy.

Abstract: Generating synthetic datasets for training face recognition models is challenging because dataset generation entails more than creating high fidelity images. It involves generating multiple images of same subjects under different factors (\textit{e.g.}, variations in pose, illumination, expression, aging and occlusion) which follows the real image conditional distribution. Previous works have studied the generation of synthetic datasets using GAN or 3D models. In this work, we approach the problem from the aspect of combining subject appearance (ID) and external factor (style) conditions. These two conditions provide a direct way to control the inter-class and intra-class variations. To this end, we propose a Dual Condition Face Generator (DCFace) based on a diffusion model. Our novel Patch-wise style extractor and Time-step dependent ID loss enables DCFace to consistently produce face images of the same subject under different styles with precise control. Face recognition models trained on synthetic images from the proposed DCFace provide higher verification accuracies compared to previous works by $6.11\%$ on average in $4$ out of $5$ test datasets, LFW, CFP-FP, CPLFW, AgeDB and CALFW. Code is available at https://github.com/mk-minchul/dcface

Abstract: Structured reconstruction is a non-trivial dense prediction problem, which extracts structural information (\eg, building corners and edges) from a raster image, then reconstructs it to a 2D planar graph accordingly. Compared with common segmentation or detection problems, it significantly relays on the capability that leveraging holistic geometric information for structural reasoning. Current transformer-based approaches tackle this challenging problem in a two-stage manner, which detect corners in the first model and classify the proposed edges (corner-pairs) in the second model. However, they separate two-stage into different models and only share the backbone encoder. Unlike the existing modeling strategies, we present an enhanced corner representation method: 1) It fuses knowledge between the corner detection and edge prediction by sharing feature in different granularity; 2) Corner candidates are proposed in four heatmap channels w.r.t its direction. Both qualitative and quantitative evaluations demonstrate that our proposed method can better reconstruct fine-grained structures, such as adjacent corners and tiny edges. Consequently, it outperforms the state-of-the-art model by +1.9\%@F-1 on Corner and +3.0\%@F-1 on Edge.

Abstract: Automatic and periodic recompiling of building databases with up-to-date high-resolution images has become a critical requirement for rapidly developing urban environments. However, the architecture of most existing approaches for change extraction attempts to learn features related to changes but ignores objectives related to buildings. This inevitably leads to the generation of significant pseudo-changes, due to factors such as seasonal changes in images and the inclination of building fa\c{c}ades. To alleviate the above-mentioned problems, we developed a contrastive learning approach by validating historical building footprints against single up-to-date remotely sensed images. This contrastive learning strategy allowed us to inject the semantics of buildings into a pipeline for the detection of changes, which is achieved by increasing the distinguishability of features of buildings from those of non-buildings. In addition, to reduce the effects of inconsistencies between historical building polygons and buildings in up-to-date images, we employed a deformable convolutional neural network to learn offsets intuitively. In summary, we formulated a multi-branch building extraction method that identifies newly constructed and removed buildings, respectively. To validate our method, we conducted comparative experiments using the public Wuhan University building change detection dataset and a more practical dataset named SI-BU that we established. Our method achieved F1 scores of 93.99% and 70.74% on the above datasets, respectively. Moreover, when the data of the public dataset were divided in the same manner as in previous related studies, our method achieved an F1 score of 94.63%, which surpasses that of the state-of-the-art method.

Abstract: This paper presents a DETR-based method for cross-domain weakly supervised object detection (CDWSOD), aiming at adapting the detector from source to target domain through weak supervision. We think DETR has strong potential for CDWSOD due to an insight: the encoder and the decoder in DETR are both based on the attention mechanism and are thus capable of aggregating semantics across the entire image. The aggregation results, i.e., image-level predictions, can naturally exploit the weak supervision for domain alignment. Such motivated, we propose DETR with additional Global Aggregation (DETR-GA), a CDWSOD detector that simultaneously makes "instance-level + image-level" predictions and utilizes "strong + weak" supervisions. The key point of DETR-GA is very simple: for the encoder / decoder, we respectively add multiple class queries / a foreground query to aggregate the semantics into image-level predictions. Our query-based aggregation has two advantages. First, in the encoder, the weakly-supervised class queries are capable of roughly locating the corresponding positions and excluding the distraction from non-relevant regions. Second, through our design, the object queries and the foreground query in the decoder share consensus on the class semantics, therefore making the strong and weak supervision mutually benefit each other for domain alignment. Extensive experiments on four popular cross-domain benchmarks show that DETR-GA significantly improves CSWSOD and advances the states of the art (e.g., 29.0% --> 79.4% mAP on PASCAL VOC --> Clipart_all dataset).

Abstract: Diffusion probabilistic models have been successful in generating high-quality and diverse images. However, traditional models, whose input and output are high-resolution images, suffer from excessive memory requirements, making them less practical for edge devices. Previous approaches for generative adversarial networks proposed a patch-based method that uses positional encoding and global content information. Nevertheless, designing a patch-based approach for diffusion probabilistic models is non-trivial. In this paper, we resent a diffusion probabilistic model that generates images on a patch-by-patch basis. We propose two conditioning methods for a patch-based generation. First, we propose position-wise conditioning using one-hot representation to ensure patches are in proper positions. Second, we propose Global Content Conditioning (GCC) to ensure patches have coherent content when concatenated together. We evaluate our model qualitatively and quantitatively on CelebA and LSUN bedroom datasets and demonstrate a moderate trade-off between maximum memory consumption and generated image quality. Specifically, when an entire image is divided into 2 x 2 patches, our proposed approach can reduce the maximum memory consumption by half while maintaining comparable image quality.

Abstract: We introduce Delta Denoising Score (DDS), a novel scoring function for text-based image editing that guides minimal modifications of an input image towards the content described in a target prompt. DDS leverages the rich generative prior of text-to-image diffusion models and can be used as a loss term in an optimization problem to steer an image towards a desired direction dictated by a text. DDS utilizes the Score Distillation Sampling (SDS) mechanism for the purpose of image editing. We show that using only SDS often produces non-detailed and blurry outputs due to noisy gradients. To address this issue, DDS uses a prompt that matches the input image to identify and remove undesired erroneous directions of SDS. Our key premise is that SDS should be zero when calculated on pairs of matched prompts and images, meaning that if the score is non-zero, its gradients can be attributed to the erroneous component of SDS. Our analysis demonstrates the competence of DDS for text based image-to-image translation. We further show that DDS can be used to train an effective zero-shot image translation model. Experimental results indicate that DDS outperforms existing methods in terms of stability and quality, highlighting its potential for real-world applications in text-based image editing.

Abstract: Visual reasoning is a long-term goal of vision research. In the last decade, several works have attempted to apply deep neural networks (DNNs) to the task of learning visual relations from images, with modest results in terms of the generalization of the relations learned. In recent years, several innovations in DNNs have been developed in order to enable learning abstract relation from images. In this work, we systematically evaluate a series of DNNs that integrate mechanism such as slot attention, recurrently guided attention, and external memory, in the simplest possible visual reasoning task: deciding whether two objects are the same or different. We found that, although some models performed better than others in generalizing the same-different relation to specific types of images, no model was able to generalize this relation across the board. We conclude that abstract visual reasoning remains largely an unresolved challenge for DNNs.

Abstract: We propose SampleDepth, a Convolutional Neural Network (CNN), that is suited for an adaptive LiDAR. Typically,LiDAR sampling strategy is pre-defined, constant and independent of the observed scene. Instead of letting a LiDAR sample the scene in this agnostic fashion, SampleDepth determines, adaptively, where it is best to sample the current frame.To do that, SampleDepth uses depth samples from previous time steps to predict a sampling mask for the current frame. Crucially, SampleDepth is trained to optimize the performance of a depth completion downstream task. SampleDepth is evaluated on two different depth completion networks and two LiDAR datasets, KITTI Depth Completion and the newly introduced synthetic dataset, SHIFT. We show that SampleDepth is effective and suitable for different depth completion downstream tasks.

Abstract: Multi-organ segmentation, which identifies and separates different organs in medical images, is a fundamental task in medical image analysis. Recently, the immense success of deep learning motivated its wide adoption in multi-organ segmentation tasks. However, due to expensive labor costs and expertise, the availability of multi-organ annotations is usually limited and hence poses a challenge in obtaining sufficient training data for deep learning-based methods. In this paper, we aim to address this issue by combining off-the-shelf single-organ segmentation models to develop a multi-organ segmentation model on the target dataset, which helps get rid of the dependence on annotated data for multi-organ segmentation. To this end, we propose a novel dual-stage method that consists of a Model Adaptation stage and a Model Ensemble stage. The first stage enhances the generalization of each off-the-shelf segmentation model on the target domain, while the second stage distills and integrates knowledge from multiple adapted single-organ segmentation models. Extensive experiments on four abdomen datasets demonstrate that our proposed method can effectively leverage off-the-shelf single-organ segmentation models to obtain a tailored model for multi-organ segmentation with high accuracy.

Abstract: Optical flow provides information on relative motion that is an important component in many computer vision pipelines. Neural networks provide high accuracy optical flow, yet their complexity is often prohibitive for application at the edge or in robots, where efficiency and latency play crucial role. To address this challenge, we build on the latest developments in event-based vision and spiking neural networks. We propose a new network architecture, inspired by Timelens, that improves the state-of-the-art self-supervised optical flow accuracy when operated both in spiking and non-spiking mode. To implement a real-time pipeline with a physical event camera, we propose a methodology for principled model simplification based on activity and latency analysis. We demonstrate high speed optical flow prediction with almost two orders of magnitude reduced complexity while maintaining the accuracy, opening the path for real-time deployments.

Abstract: Point clouds are widely regarded as one of the best dataset types for urban mapping purposes. Hence, point cloud datasets are commonly investigated as benchmark types for various urban interpretation methods. Yet, few researchers have addressed the use of point cloud benchmarks for fa\c{c}ade segmentation. Robust fa\c{c}ade segmentation is becoming a key factor in various applications ranging from simulating autonomous driving functions to preserving cultural heritage. In this work, we present a method of enriching existing point cloud datasets with fa\c{c}ade-related classes that have been designed to facilitate fa\c{c}ade segmentation testing. We propose how to efficiently extend existing datasets and comprehensively assess their potential for fa\c{c}ade segmentation. We use the method to create the TUM-FA\c{C}ADE dataset, which extends the capabilities of TUM-MLS-2016. Not only can TUM-FA\c{C}ADE facilitate the development of point-cloud-based fa\c{c}ade segmentation tasks, but our procedure can also be applied to enrich further datasets.

Abstract: For visual estimation of optical flow, a crucial function for many vision tasks, unsupervised learning, using the supervision of view synthesis has emerged as a promising alternative to supervised methods, since ground-truth flow is not readily available in many cases. However, unsupervised learning is likely to be unstable when pixel tracking is lost due to occlusion and motion blur, or the pixel matching is impaired due to variation in image content and spatial structure over time. In natural environments, dynamic occlusion or object variation is a relatively slow temporal process spanning several frames. We, therefore, explore the optical flow estimation from multiple-frame sequences of dynamic scenes, whereas most of the existing unsupervised approaches are based on temporal static models. We handle the unsupervised optical flow estimation with a temporal dynamic model by introducing a spatial-temporal dual recurrent block based on the predictive coding structure, which feeds the previous high-level motion prior to the current optical flow estimator. Assuming temporal smoothness of optical flow, we use motion priors of the adjacent frames to provide more reliable supervision of the occluded regions. To grasp the essence of challenging scenes, we simulate various scenarios across long sequences, including dynamic occlusion, content variation, and spatial variation, and adopt self-supervised distillation to make the model understand the object's motion patterns in a prolonged dynamic environment. Experiments on KITTI 2012, KITTI 2015, Sintel Clean, and Sintel Final datasets demonstrate the effectiveness of our methods on unsupervised optical flow estimation. The proposal achieves state-of-the-art performance with advantages in memory overhead.

Abstract: Solar activity is one of the main drivers of variability in our solar system and the key source of space weather phenomena that affect Earth and near Earth space. The extensive record of high resolution extreme ultraviolet (EUV) observations from the Solar Dynamics Observatory (SDO) offers an unprecedented, very large dataset of solar images. In this work, we make use of this comprehensive dataset to investigate capabilities of current state-of-the-art generative models to accurately capture the data distribution behind the observed solar activity states. Starting from StyleGAN-based methods, we uncover severe deficits of this model family in handling fine-scale details of solar images when training on high resolution samples, contrary to training on natural face images. When switching to the diffusion based generative model family, we observe strong improvements of fine-scale detail generation. For the GAN family, we are able to achieve similar improvements in fine-scale generation when turning to ProjectedGANs, which uses multi-scale discriminators with a pre-trained frozen feature extractor. We conduct ablation studies to clarify mechanisms responsible for proper fine-scale handling. Using distributed training on supercomputers, we are able to train generative models for up to 1024x1024 resolution that produce high quality samples indistinguishable to human experts, as suggested by the evaluation we conduct. We make all code, models and workflows used in this study publicly available at \url{https://github.com/SLAMPAI/generative-models-for-highres-solar-images}.

Abstract: In recent years, media reports have called out bias and racism in face recognition technology. We review experimental results exploring several speculated causes for asymmetric cross-demographic performance. We consider accuracy differences as represented by variations in non-mated (impostor) and / or mated (genuine) distributions for 1-to-1 face matching. Possible causes explored include differences in skin tone, face size and shape, imbalance in number of identities and images in the training data, and amount of face visible in the test data ("face pixels"). We find that demographic differences in face pixel information of the test images appear to most directly impact the resultant differences in face recognition accuracy.

Abstract: The recent breakthroughs in natural language processing for model pretraining on large quantities of data have opened the way for similar foundation models in computer vision. These models could greatly simplify the use of images in any system by producing all-purpose visual features, i.e., features that work across image distributions and tasks without finetuning. This work shows that existing pretraining methods, especially self-supervised methods, can produce such features if trained on enough curated data from diverse sources. We revisit existing approaches and combine different techniques to scale our pretraining in terms of data and model size. Most of the technical contributions aim at accelerating and stabilizing the training at scale. In terms of data, we propose an automatic pipeline to build a dedicated, diverse, and curated image dataset instead of uncurated data, as typically done in the self-supervised literature. In terms of models, we train a ViT model (Dosovitskiy et al., 2020) with 1B parameters and distill it into a series of smaller models that surpass the best available all-purpose features, OpenCLIP (Ilharco et al., 2021) on most of the benchmarks at image and pixel levels.

Abstract: Understanding semantic scene segmentation of urban scenes captured from the Unmanned Aerial Vehicles (UAV) perspective plays a vital role in building a perception model for UAV. With the limitations of large-scale densely labeled data, semantic scene segmentation for UAV views requires a broad understanding of an object from both its top and side views. Adapting from well-annotated autonomous driving data to unlabeled UAV data is challenging due to the cross-view differences between the two data types. Our work proposes a novel Cross-View Adaptation (CROVIA) approach to effectively adapt the knowledge learned from on-road vehicle views to UAV views. First, a novel geometry-based constraint to cross-view adaptation is introduced based on the geometry correlation between views. Second, cross-view correlations from image space are effectively transferred to segmentation space without any requirement of paired on-road and UAV view data via a new Geometry-Constraint Cross-View (GeiCo) loss. Third, the multi-modal bijective networks are introduced to enforce the global structural modeling across views. Experimental results on new cross-view adaptation benchmarks introduced in this work, i.e., SYNTHIA to UAVID and GTA5 to UAVID, show the State-of-the-Art (SOTA) performance of our approach over prior adaptation methods

Abstract: Vegetation structure mapping is critical for understanding the global carbon cycle and monitoring nature-based approaches to climate adaptation and mitigation. Repeat measurements of these data allow for the observation of deforestation or degradation of existing forests, natural forest regeneration, and the implementation of sustainable agricultural practices like agroforestry. Assessments of tree canopy height and crown projected area at a high spatial resolution are also important for monitoring carbon fluxes and assessing tree-based land uses, since forest structures can be highly spatially heterogeneous, especially in agroforestry systems. Very high resolution satellite imagery (less than one meter (1m) ground sample distance) makes it possible to extract information at the tree level while allowing monitoring at a very large scale. This paper presents the first high-resolution canopy height map concurrently produced for multiple sub-national jurisdictions. Specifically, we produce canopy height maps for the states of California and S\~{a}o Paolo, at sub-meter resolution, a significant improvement over the ten meter (10m) resolution of previous Sentinel / GEDI based worldwide maps of canopy height. The maps are generated by applying a vision transformer to features extracted from a self-supervised model in Maxar imagery from 2017 to 2020, and are trained against aerial lidar and GEDI observations. We evaluate the proposed maps with set-aside validation lidar data as well as by comparing with other remotely sensed maps and field-collected data, and find our model produces an average Mean Absolute Error (MAE) within set-aside validation areas of 3.0 meters.

Abstract: Recently, pre-trained point cloud models have found extensive applications in downstream tasks like object classification. However, these tasks often require {full fine-tuning} of models and lead to storage-intensive procedures, thus limiting the real applications of pre-trained models. Inspired by the great success of visual prompt tuning (VPT) in vision, we attempt to explore prompt tuning, which serves as an efficient alternative to full fine-tuning for large-scale models, to point cloud pre-trained models to reduce storage costs. However, it is non-trivial to apply the traditional static VPT to point clouds, owing to the distribution diversity of point cloud data. For instance, the scanned point clouds exhibit various types of missing or noisy points. To address this issue, we propose an Instance-aware Dynamic Prompt Tuning (IDPT) for point cloud pre-trained models, which utilizes a prompt module to perceive the semantic prior features of each instance. This semantic prior facilitates the learning of unique prompts for each instance, thus enabling downstream tasks to robustly adapt to pre-trained point cloud models. Notably, extensive experiments conducted on downstream tasks demonstrate that IDPT outperforms full fine-tuning in most tasks with a mere 7\% of the trainable parameters, thus significantly reducing the storage pressure. Code is available at \url{https://github.com/zyh16143998882/IDPT}.

Abstract: Pedestrian attribute recognition (PAR) has received increasing attention because of its wide application in video surveillance and pedestrian analysis. Extracting robust feature representation is one of the key challenges in this task. The existing methods mainly use the convolutional neural network (CNN) as the backbone network to extract features. However, these methods mainly focus on small discriminative regions while ignoring the global perspective. To overcome these limitations, we propose a pure transformer-based multi-task PAR network named PARFormer, which includes four modules. In the feature extraction module, we build a transformer-based strong baseline for feature extraction, which achieves competitive results on several PAR benchmarks compared with the existing CNN-based baseline methods. In the feature processing module, we propose an effective data augmentation strategy named batch random mask (BRM) block to reinforce the attentive feature learning of random patches. Furthermore, we propose a multi-attribute center loss (MACL) to enhance the inter-attribute discriminability in the feature representations. In the viewpoint perception module, we explore the impact of viewpoints on pedestrian attributes, and propose a multi-view contrastive loss (MCVL) that enables the network to exploit the viewpoint information. In the attribute recognition module, we alleviate the negative-positive imbalance problem to generate the attribute predictions. The above modules interact and jointly learn a highly discriminative feature space, and supervise the generation of the final features. Extensive experimental results show that the proposed PARFormer network performs well compared to the state-of-the-art methods on several public datasets, including PETA, RAP, and PA100K. Code will be released at https://github.com/xwf199/PARFormer.

Abstract: The localization of objects is a crucial task in various applications such as robotics, virtual and augmented reality, and the transportation of goods in warehouses. Recent advances in deep learning have enabled the localization using monocular visual cameras. While structure from motion (SfM) predicts the absolute pose from a point cloud, absolute pose regression (APR) methods learn a semantic understanding of the environment through neural networks. However, both fields face challenges caused by the environment such as motion blur, lighting changes, repetitive patterns, and feature-less structures. This study aims to address these challenges by incorporating additional information and regularizing the absolute pose using relative pose regression (RPR) methods. The optical flow between consecutive images is computed using the Lucas-Kanade algorithm, and the relative pose is predicted using an auxiliary small recurrent convolutional network. The fusion of absolute and relative poses is a complex task due to the mismatch between the global and local coordinate systems. State-of-the-art methods fusing absolute and relative poses use pose graph optimization (PGO) to regularize the absolute pose predictions using relative poses. In this work, we propose recurrent fusion networks to optimally align absolute and relative pose predictions to improve the absolute pose prediction. We evaluate eight different recurrent units and construct a simulation environment to pre-train the APR and RPR networks for better generalized training. Additionally, we record a large database of different scenarios in a challenging large-scale indoor environment that mimics a warehouse with transportation robots. We conduct hyperparameter searches and experiments to show the effectiveness of our recurrent fusion method compared to PGO.

Abstract: Object detection has seen remarkable progress in recent years with the introduction of Convolutional Neural Networks (CNN). Object detection is a multi-task learning problem where both the position of the objects in the images as well as their classes needs to be correctly identified. The idea here is to maximize the overlap between the ground-truth bounding boxes and the predictions i.e. the Intersection over Union (IoU). In the scope of work seen currently in this domain, IoU is approximated by using the Huber loss as a proxy but this indirect method does not leverage the IoU information and treats the bounding box as four independent, unrelated terms of regression. This is not true for a bounding box where the four coordinates are highly correlated and hold a semantic meaning when taken together. The direct optimization of the IoU is not possible due to its non-convex and non-differentiable nature. In this paper, we have formulated a novel loss namely, the Smooth IoU, which directly optimizes the IoUs for the bounding boxes. This loss has been evaluated on the Oxford IIIT Pets, Udacity self-driving car, PASCAL VOC, and VWFS Car Damage datasets and has shown performance gains over the standard Huber loss.

Abstract: Domain generalization (DG), aiming at models able to work on multiple unseen domains, is a must-have characteristic of general artificial intelligence. DG based on single source domain training data is more challenging due to the lack of comparable information to help identify domain invariant features. In this paper, it is determined that the domain invariant features could be contained in the single source domain training samples, then the task is to find proper ways to extract such domain invariant features from the single source domain samples. An assumption is made that the domain invariant features are closely related to the frequency. Then, a new method that learns through multiple frequency domains is proposed. The key idea is, dividing the frequency domain of each original image into multiple subdomains, and learning features in the subdomain by a designed two branches network. In this way, the model is enforced to learn features from more samples of the specifically limited spectrum, which increases the possibility of obtaining the domain invariant features that might have previously been defiladed by easily learned features. Extensive experimental investigation reveals that 1) frequency decomposition can help the model learn features that are difficult to learn. 2) the proposed method outperforms the state-of-the-art methods of single-source domain generalization.

Abstract: The strength of machine learning models stems from their ability to learn complex function approximations from data; however, this strength also makes training deep neural networks challenging. Notably, the complex models tend to memorize the training data, which results in poor regularization performance on test data. The regularization techniques such as L1, L2, dropout, etc. are proposed to reduce the overfitting effect; however, they bring in additional hyperparameters tuning complexity. These methods also fall short when the inter-class similarity is high due to the underlying data distribution, leading to a less accurate model. In this paper, we present a novel approach to regularize the models by leveraging the information-rich latent embeddings and their high intra-class correlation. We create phantom embeddings from a subset of homogenous samples and use these phantom embeddings to decrease the inter-class similarity of instances in their latent embedding space. The resulting models generalize better as a combination of their embedding and regularize them without requiring an expensive hyperparameter search. We evaluate our method on two popular and challenging image classification datasets (CIFAR and FashionMNIST) and show how our approach outperforms the standard baselines while displaying better training behavior.