WACV 2024 Papers

Scene understanding plays an essential role in enabling autonomous driving and maintaining high standards of performance and safety. To address this task, cameras and laser scanners (LiDARs) have been the most commonly used sensors, with radars being less popular. Despite that, radars remain low-cost, information-dense, and fast-sensing techniques that are resistant to adverse weather conditions. While multiple works have been previously presented for radar-based scene semantic segmentation, the nature of the radar data still poses a challenge due to the inherent noise and sparsity, as well as the disproportionate foreground and background. In this work, we propose a novel approach to the semantic segmentation of radar scenes using a multi-input fusion of radar data through a novel architecture and loss functions that are tailored to tackle the drawbacks of radar perception. Our novel architecture includes an efficient attention block that adaptively captures important feature information. Our method, TransRadar, outperforms state-of-the-art methods on the CARRADA and RADIal datasets while having smaller model sizes.

Federated learning has emerged as a promising paradigm for collaborative machine learning, enabling multiple clients to train a model while preserving data privacy jointly. Tailored federated learning takes this concept further by accommodating client heterogeneity and facilitating the learning of personalized models. While the utilization of transformers within federated learning has attracted significant interest, there remains a need to investigate the effects of federated learning algorithms on the latest focal modulation-based transformers. In this paper, we investigate this relationship and uncover the detrimental effects of federated averaging (FedAvg) algorithms on Focal Modulation, particularly in scenarios with heterogeneous data. To address this challenge, we propose TransFed, a novel transformer-based federated learning framework that not only aggregates model parameters but also learns tailored Focal Modulation for each client. Instead of employing a conventional customization mechanism that maintains client-specific focal modulation layers locally, we introduce a learn-to-tailor approach that fosters client collaboration, enhancing scalability and adaptation in TransFed. Our method incorporates a hyper network on the server, responsible for learning personalized projection matrices for the focal modulation layers. This enables the generation of client-specific keys, values, and queries. Furthermore, we provide an analysis of adaptation bounds for TransFed using the learn-to-customize mechanism. Through intensive experiments on datasets related to pneumonia classification, we demonstrate that TransFed, in combination with the learn-to-tailor approach, achieves superior performance in scenarios with non-IID data distributions, surpassing existing methods. Overall, TransFed paves the way for leveraging focal Modulation in federated learning, advancing the capabilities of focal modulated transformer models in decentralized environments.

This paper tackles the problem of object counting in images. Existing approaches rely on extensive training data with point annotations for each object, making data collection labor-intensive and time-consuming. To overcome this, we propose a training-free object counter that treats the counting task as a segmentation problem. Our approach leverages the Segment Anything Model (SAM), known for its high-quality masks and zero-shot segmentation capability. However, the vanilla mask generation method of SAM lacks class-specific information in the masks, resulting in inferior counting accuracy. To overcome this limitation, we introduce a prior-guided mask generation method that incorporates three types of priors into the segmentation process, enhancing efficiency and accuracy. Additionally, we tackle the issue of counting objects specified through text by proposing a two-stage approach that combines reference object selection and prior-guided mask generation. Extensive experiments on standard datasets demonstrate the competitive performance of our training-free counter compared to learning-based approaches. This paper presents a promising solution for counting objects in various scenarios without the need for extensive data collection and counting-specific training. Code is available at https://github.com/shizenglin/training-free-object-counter.

Recent diffusion-based generators can produce high-quality images from textual prompts. However, they often disregard textual instructions that specify the spatial layout of the composition. We propose a simple approach that achieves robust layout control without the need for training or fine-tuning of the image generator. Our technique manipulates the cross-attention layers that the model uses to interface textual and visual information and steers the generation in the desired direction given, e.g., a user-specified layout. To determine how to best guide attention, we study the role of attention maps and explore two alternative strategies, forward and backward guidance. We thoroughly evaluate our approach on three benchmarks and provide several qualitative examples and a comparative analysis of the two strategies that demonstrate the superiority of backward guidance compared to forward guidance, as well as prior work. We further demonstrate the versatility of layout guidance by extending it to applications such as editing the layout and context of real images.

Diffusion models (DMs) synthesize high-quality images in various domains. However, controlling their generative process is still hazy because the intermediate variables in the process are not rigorously studied. Recently, the bottleneck feature of the U-Net, namely h-space, is found to convey the semantics of the resulting image. It enables StyleCLIP-like latent editing within DMs. In this paper, we explore further usage of h-space beyond attribute editing, and introduce a method to inject the content of one image into another image by combining their features in the generative processes. Briefly, given the original generative process of the other image, 1) we gradually blend the bottleneck feature of the content with proper normalization, and 2) we calibrate the skip connections to match the injected content. Unlike custom-diffusion approaches, our method does not require time-consuming optimization or fine-tuning. Instead, our method manipulates intermediate features within a feed-forward generative process. Furthermore, our method does not require supervision from external networks.

Semi-supervised object detection (SSOD) aims to improve the performance and generalization of existing object detectors by utilizing limited labeled data and extensive unlabeled data. Despite many advances, recent SSOD methods are still challenged by inadequate model refinement using the classical exponential moving average (EMA) strategy, the consensus of Teacher-Student models in the latter stages of training (i.e., losing their distinctiveness), and noisy/misleading pseudo-labels. This paper proposes a novel training-based model refinement (TMR) stage and a simple yet effective representation disagreement (RD) strategy to address the limitations of classical EMA and the consensus problem. The TMR stage of Teacher-Student models optimizes the lightweight scaling operation to refine the model's weights and prevent overfitting or forgetting learned patterns from unlabeled data. Meanwhile, the RD strategy helps keep these models diverged to encourage the student model to explore additional patterns in unlabeled data. Our approach can be integrated into established SSOD methods and is empirically validated using two baseline methods, with and without cascade regression, to generate more reliable pseudo-labels. Extensive experiments demonstrate the superior performance of our approach over state-of-the-art SSOD methods. Specifically, the proposed approach outperforms the baseline Unbiased-Teacher-v2 (& Unbiased-Teacher-v1) method by an average mAP margin of 2.23, 2.1, and 3.36 (& 2.07, 1.9, and 3.27) on COCO-standard, COCO-additional, and Pascal VOC datasets, respectively.

Deep active learning in the presence of outlier examples poses a realistic yet challenging scenario. Acquiring unlabeled data for annotation requires a delicate balance between avoiding outliers to conserve the annotation budget and prioritizing useful inlier examples for effective training. In this work, we present an approach that leverages three highly synergistic components, which are identified as key ingredients: joint classifier training with inliers and outliers, semi-supervised learning through pseudo-labeling, and model ensembling. Our work demonstrates that ensembling significantly enhances the accuracy of pseudo-labeling and improves the quality of data acquisition. By enabling semi-supervision through the joint training process, where outliers are properly handled, we observe a substantial boost in classifier accuracy through the use of all available unlabeled examples. Notably, we reveal that the integration of joint training renders explicit outlier detection unnecessary; a conventional component for acquisition in prior work. The three key components align seamlessly with numerous existing approaches. Through empirical evaluations, we showcase that their combined use leads to a performance increase. Remarkably, despite its simplicity, our proposed approach outperforms all other methods in terms of performance. Code: https://github.com/vladan-stojnic/active-outliers

Visual tracking of tiny and low-contrast objects such as insects in cluttered natural environments is a very challenging computer vision task. This is particularly true for machine learning algorithms, which usually require distinct visual foreground features to reliably identify the object of interest. Here, we propose a novel deep learning-based tracking framework capable of detecting tiny and visually camouflaged ants (covering only a few pixels) in complex and dynamic high-resolution videos. In particular, we introduce refinable recurrent Hourglass Networks, which combine colour and temporal information to continuously detect insects recorded using a freely moving camera. Moreover, this architecture provides comprehensible heatmaps of positional estimations and a seamless integration of optional user-input to further refine the tracking results if necessary. We evaluated our algorithm on an extremely challenging wildlife ant dataset with a resolution of 1024x1024 and report a mean deviation of 19 pixels from the ground truth (object 30 px) without any user input. By providing only 0.6% manual locations this accuracy can be improved to a mean deviation of 9 pixels. A comparison to a well known deep learning-based single frame detection algorithm (YOLOv7), two state-of-the-art tracking methods (ToMP and KeepTrack), a probabilistic tracking framework and a comprehensive ablation study reveal superior performances in all our experiments. Our tracking framework therefore provides a foundation for challenging tiny single-object tracking scenarios and a practical and interactive solution for biologists and ecologists.

Skiing is a popular winter sport discipline with a long history of competitive events. In this domain, computer vision has the potential to enhance the understanding of athletes' performance, but its application lags behind other sports due to limited studies and datasets. This paper makes a step forward in filling such gaps. A thorough investigation is performed on the task of skier tracking in a video capturing his/her complete performance. Obtaining continuous and accurate skier localization is preemptive for further higher-level performance analyses. To enable the study, the largest and most annotated dataset for computer vision in skiing, SkiTB, is introduced. Several visual object tracking algorithms, including both established methodologies and a newly introduced skier-optimized baseline algorithm, are tested using the dataset. The results provide valuable insights into the applicability of different tracking methods for vision-based skiing analysis. SkiTB, code, and results are available at https://machinelearning.uniud.it/datasets/skitb.

This paper introduces a visual real-time insect monitoring approach capable of detecting and tracking tiny and fast-moving objects in cluttered wildlife conditions using an RGB-DVS stereo-camera system. By building on the intrinsic benefits of event vision data acquisition, we demonstrate that insect presence can be detected at an extremely high temporal rate (on average more than 40 times real-time) while surpassing the spatial and spectral sensitivity of conventional colour-based sensing. Our DVS-based detection and tracking algorithm extracts insect locations over time, and we evaluated our system based on 81104 manually annotated stereo-frames with 34453 insect appearances featuring highly varying scenes and imaging conditions (including clutter, wind-induced motion, etc.). Comparing our algorithm to two state-of-the-art deep learning algorithms reveals superior results in both detection performance and computational speed. Using the DVS as a trigger for the temporally synchronised RGB camera, we are able to correctly identify 73% of images with and without insects which can be increased to 76% with parameters optimised for different scenes. Overall, our study suggests that DVS-based sensing can be used for visual insect monitoring by enabling reliable real-time insect detection in wildlife conditions while significantly reducing the necessity for data storage, manual labour and energy.

In the past years, deep convolutional neural networks have been pushing the frontier of face recognition (FR) techniques in both verification and identification scenarios. Despite the high accuracy, they are often criticized for lacking explainability. There has been an increasing demand for understanding the decision-making process of deep face recognition systems. Recent studies have investigated the usage of visual saliency maps as an explanation, but they often lack a discussion and analysis in the context of face recognition. This paper concentrates on explainable face verification tasks and conceives a new explanation framework. Firstly, a definition of the saliency-based explanation method is provided, which focuses on the decisions made by the deep FR model. Secondly, a new model-agnostic explanation method named CorrRISE is proposed to produce saliency maps, which reveal both the similar and dissimilar regions of any given pair of face images. Then, an evaluation methodology is designed to measure the performance of general visual saliency explanation methods in face verification. Finally, substantial visual and quantitative results have shown that the proposed CorrRISE method demonstrates promising results in comparison with other state-of-the-art explainable face verification approaches.

In recent years, there has been significant progress in 2D generative face models fueled by applications such as animation, synthetic data generation, and digital avatars. However, due to the absence of 3D information, these 2D models often struggle to accurately disentangle facial attributes like pose, expression, and illumination, limiting their editing capabilities. To address this limitation, this paper proposes a 3D controllable generative face model to produce high-quality albedo and precise 3D shapes by leveraging existing 2D generative models. By combining 2D face generative models with semantic face manipulation, this method enables editing of detailed 3D rendered faces. The proposed framework utilizes an alternating descent optimization approach over shape and albedo. Differentiable rendering is used to train high-quality shapes and albedo without 3D supervision. Moreover, this approach outperforms most state-of-the-art (SOTA) methods in the well-known NoW and REALY benchmarks for 3D face reconstruction. It also outperforms the SOTA reconstruction models in recovering rendered faces' identities across novel poses. Additionally, the paper demonstrates direct control of expressions in 3D faces by exploiting latent space leading to text-based editing of 3D faces.

Generative diffusion models, including Stable Diffusion and Midjourney, can generate visually appealing, diverse, and high-resolution images for various applications. These models are trained on billions of internet-sourced images, raising significant concerns about the potential unauthorized use of copyright-protected images. In this paper, we examine whether it is possible to determine if a specific image was used in the training set, a problem known as a membership inference attack. Our focus is on Stable Diffusion, and we address the challenge of designing a fair evaluation framework to answer this membership question. We propose a new dataset to establish a fair evaluation setup and apply it to Stable Diffusion, also applicable to other generative models. With the proposed dataset, we execute membership attacks (both known and newly introduced). Our research reveals that previously proposed evaluation setups do not provide a full understanding of the effectiveness of membership inference attacks. We conclude that the membership inference attack remains a significant challenge for large diffusion models (often deployed as black-box systems), indicating that related privacy and copyright issues will persist in the foreseeable future.

Generalization on unseen domains is critical for Deep Neural Networks (DNNs) to perform well in real-world applications such as autonomous navigation. However, catastrophic forgetting limit the ability of domain generalization and unsupervised domain adaption approaches to adapt to constantly changing target domains. To overcome these challenges, We propose DoSe framework, a Domain-aware Self-Distillation method based on batch normalization prototypes to facilitate continual model generalization across varying target domains. Specifically, we enforce the consistency of batch normalization statistics between two batches of images sampled from the same target domain distribution between the student and teacher models. To alleviate catastrophic forgetting, we introduce a novel exemplar-based replay buffer to identify difficult samples for the model to retain the knowledge. Specifically, we demonstrate that identifying difficult samples and updating the model periodically using them can help in preserving knowledge learned from previously seen domains. We conduct extensive experiments on two real-world datasets ACDC, C-Driving, and one synthetic dataset SHIFT to verify the efficiency of the proposed DoSe framework. On ACDC, our method outperforms existing SOTA in Domain Generalization, Unsupervised Domain Adaptation, and Daytime settings by 26%, 14%, and 70% respectively.

In this work, we consider the typography generation task that aims at producing diverse typographic styling for the given graphic document. We formulate typography generation as a fine-grained attribute generation for multiple text elements and build an autoregressive model to generate diverse typography that matches the input design context. We further propose a simple yet effective sampling approach that respects the consistency and distinction principle of typography so that generated examples share consistent typographic styling across text elements. Our empirical study shows that our model successfully generates diverse typographic designs while preserving a consistent typographic structure.

We present SPSRC, a novel, simple and effective framework to extract enhanced Structured Pruning Saliency scores by Reorganizing Convolution. We observe that performance of pruning methods have gradually plateaued recently and propose to make better use of the learned convolutional kernel weights simply after a few steps of transformations. We firstly re-organize the convolutional operations between layers as matrix multiplications and then use the singular values and the matrix norms of the transformed matrices as saliency scores to decide what channels to prune or keep. We show both analytically and empirically that the long-standing kernel-norm-based channel importance measurement, like L1 magnitude, is not precise enough possessing a very obvious weakness of lacking spatial saliency but can be improved with SPSRC. We conduct extensive experiments across different settings and configurations and compare with the counterparts without our SPSRC along with other popular methods, observing obvious improvements. Our code is available at https://github.com/AlexSunNik/SPSRC/tree/main.

Object proposal generation serves as a standard pre-processing step in Vision-Language (VL) tasks (image captioning, visual question answering, etc.). The performance of object proposals generated for VL tasks is currently evaluated across all available annotations, a protocol that we show is misaligned - higher scores do not necessarily correspond to improved performance on downstream VL tasks. Our work serves as a study of this phenomenon and explores the effectiveness of semantic grounding to mitigate its effects. To this end, we propose evaluating object proposals against only a subset of available annotations, selected by thresholding an annotation importance score. Importance of object annotations to VL tasks is quantified by extracting relevant semantic information from text describing the image. We show that our method is consistent and demonstrates greatly improved alignment with annotations selected by image captioning metrics and human annotation when compared against existing techniques. Lastly, we compare current detectors used in the Scene Graph Generation (SGG) benchmark as a use case, which serves as an example of when traditional object proposal evaluation techniques are misaligned.

Automated disease segmentation in plant images plays a crucial role in identifying and mitigating the impact of plant diseases on agricultural productivity. In this study, we address the problem of Northern Leaf Blight (NLB) disease segmentation in maize plants. We present a comprehensive dataset of 1000 plant images annotated with NLB disease regions. We employ the Mask R-CNN and Cascaded Mask R-CNN models with various backbone architectures to perform NLB disease segmentation. The experimental results demonstrate the effectiveness of the models in accurately delineating NLB disease regions. Specifically, the ResNet Strikes Back-50 backbone architecture achieves the highest mean average precision (mAP) score, indicating its ability to capture intricate details of NLB disease spots. Additionally, the cascaded approach enhances segmentation accuracy compared to the single-stage Mask R-CNN models. Our findings provide valuable insights into the performance of different backbone architectures and contribute to the development of automated NLB disease segmentation methods in plant images. The generated dataset and experimental results serve as a resource for further research in plant disease segmentation and management.

Despite the widespread deployment of outdoor cameras, their potential for automated analysis remains largely untapped due, in part, to calibration challenges. The absence of precise camera calibration data, including intrinsic and extrinsic parameters, hinders accurate real-world distance measurements from captured videos. To address this, we present a scalable framework that utilizes street-level imagery to reconstruct a metric 3D model, facilitating precise calibration of in-the-wild traffic cameras. Notably, our framework achieves 3D scene reconstruction and accurate localization of over 100 global traffic cameras and is scalable to any camera with sufficient street-level imagery. For evaluation, we introduce a dataset of 20 fully calibrated traffic cameras, demonstrating our method's significant enhancements over existing automatic calibration techniques. Furthermore, we highlight our approach's utility in traffic analysis by extracting insights via 3D vehicle reconstruction and speed measurement, thereby opening up the potential of using outdoor cameras for automated analysis.