WACV 2024 论文列表

Infertility is a global health problem, and an increasing number of couples are seeking medical assistance to achieve reproduction, at least half of which are caused by men. The success rate of assisted reproductive technologies depends on sperm assessment, in which experts determine whether sperm can be used for reproduction based on morphology and motility of sperm. Previous sperm assessment studies with deep learning have used datasets comprising images that include only sperm heads, which cannot consider motility and other morphologies of sperm. Furthermore, the labels of the dataset are one-hot, which provides insufficient support for experts, because assessment results are inconsistent between experts, and they have no absolute answer. Therefore, we constructed the video dataset for sperm assessment whose videos include sperm head as well as neck and tail, and its labels were annotated with soft-label. Furthermore, we proposed the sperm assessment framework and the neural network, RoSTFine, for sperm video recognition. Experimental results showed that RoSTFine could improve the sperm assessment performances compared to existing video recognition models and focus strongly on important sperm parts (i.e., head and neck).

We propose a system for automated monitoring of ear-biting in pigs. Ear-biting presents a welfare challenge to commercial pig farming, leading to injuries and infections that affect animal welfare. We use a computer vision system to detect and track all pigs and ear-biting events. Our goal is to provide early warning of ear-biting to allow quick intervention to improve the health and welfare of commercial farm animals. We compare several different object detection methods for the detection of individual pigs, including an oriented bounding box detector, which is better suited to the accurate detection of pigs from overhead cameras. We track all pigs and all ear-biting events using a specialised two-stage multi-object tracking system. The tracking system is adapted to match the characteristics of each entity being tracked. The tracking system allows the individual pigs involved in an ear-biting incident to be identified, allowing for targeted welfare interventions. We evaluate our complete system on real farm videos and demonstrate that our complete system improves compared to existing ear-biting detection methods.

Over the past few decades, a significant rise of camera-based applications for traffic monitoring has occurred. Governments and local administrations are increasingly relying on the data collected from these cameras to enhance road safety and optimize traffic conditions. However, for effective data utilization, it is imperative to ensure accurate and automated calibration of the involved cameras. This paper proposes a novel approach to address this challenge by leveraging the topological structure of intersections. We propose a framework involving the generation of a set of synthetic intersection viewpoint images from a bird's-eye-view image, framed as a graph of virtual cameras to model these images. Using the capabilities of Graph Neural Networks, we effectively learn the relationships within this graph, thereby facilitating the estimation of a homography matrix. This estimation leverages the neighbourhood representation for any real-world camera and is enhanced by exploiting multiple images instead of a single match. In turn, the homography matrix allows the retrieval of extrinsic calibration parameters. As a result, the proposed framework demonstrates superior performance on both synthetic datasets and real-world cameras, setting a new state-of-the-art benchmark.

The Ball-Pivoting Algorithm (BPA) is a notable technique for 3D surface reconstruction from point clouds, heavily reliant on the ball radius. In practical application, determining the optimal radius for BPA often necessitates iterative experimentation to achieve better reconstruction quality. BPA entails geometric computations like iterative pivoting, inherently lacking differentiability. In this paper, we tackle the dual challenges of radius selection and non-differentiability in BPA. Inspired by contextual bandits, we propose an innovative approach that learns the optimal radius based on local geometric features within point clouds. We validate our method on the ModelNet10 and ShapeNet datasets, showcasing superior surface reconstruction compared to manual tuning and other classic methods both for low and high point cloud densities. Our code is available at https://github.com/houda-pixel/Auto-BPA.

Grounding-based vision and language models have been successfully applied to low-level vision tasks, aiming to precisely locate objects referred in captions. The effectiveness of grounding representation learning heavily relies on the scale of the training dataset. Despite being a useful data enrichment strategy, data augmentation has received minimal attention in existing vision and language tasks as augmentation for image-caption pairs is non-trivial. In this study, we propose a robust phrase grounding model trained with text-conditioned and text-unconditioned data augmentations. Specifically, we apply text-conditioned color jittering and horizontal flipping to ensure semantic consistency between images and captions. To guarantee image-caption correspondence in the training samples, we modify the captions according to pre-defined keywords when applying horizontal flipping. Additionally, inspired by recent masked signal reconstruction, we propose to use pixel-level masking as a novel form of data augmentation. While we demonstrate our data augmentation method with MDETR framework, the proposed approach is applicable to common grounding-based vision and language tasks with other frameworks. Finally, we show that larger capacity image encoder such as CLIP can further improve the results. Through extensive experiments on three commonly applied datasets: Flickr30k, referring expressions, and GQA, our method demonstrates advanced performance over the state-of-the-arts with various metrics. Code can be found in https://github.com/amzn/augment-the-pairs-wacv2024.

3D object detection networks tend to be biased towards the data they are trained on. Evaluation on datasets captured in different locations, conditions or sensors than that of the training (source) data results in a drop in model performance due to the gap in distribution with the test (or target) data. Current methods for domain adaptation either assume access to source data during training, which may not be available due to privacy or memory concerns, or require a sequence of lidar frames as an input. We propose a single-frame approach for source-free, unsupervised domain adaptation of lidar-based 3D object detectors that uses class prototypes to mitigate the effect pseudo-label noise. Addressing the limitations of traditional feature aggregation methods for prototype computation in the presence of noisy labels, we utilize a transformer module to identify outlier ROI's that correspond to incorrect, over-confident annotations, and compute an attentive class prototype. Under an iterative training strategy, the losses associated with noisy pseudo labels are down-weighed and thus refined in the process of self-training. To validate the effectiveness of our proposed approach, we examine the domain shift associated with networks trained on large, label-rich datasets (such as the Waymo Open Dataset and nuScenes) and evaluate on smaller, label-poor datasets (such as KITTI) and vice-versa. We demonstrate our approach on two recent object detectors and achieve results that out-perform the other domain adaptation works.

Real-world medical datasets often suffer from class imbalance, which can lead to degraded performance due to limited samples of the minority class. In another line of research, Transformer-based multiple instance learning (Transformer-MIL) has shown promise in addressing the pairwise correlation between instances in medical whole slide images (WSIs) with gigapixel resolution and non-uniform sizes. However, these characteristics pose challenges for state-of-the-art (SOTA) oversampling methods aiming at diversifying the minority context in imbalanced WSIs. In this paper, we propose an Attention-Guided Prototype Mixing scheme at the WSI level. We leverage Transformer-MIL training to determine the distribution of semantic instances and identify relevant instances for cutting and pasting across different WSI (bag of instances). To our knowledge, applying Transformer is often limited by memory requirements and time complexity, particularly when dealing with gigabyte-sized WSIs. We introduce the concept of prototype instances that have smaller representations while preserving the uniform size and intrinsic features of the WSI. We demonstrate that our proposed method can boost performance compared to competitive SOTA oversampling and augmentation methods at an imbalanced WSI level.

Within (semi-)automated visual industrial inspection, learning-based approaches for assessing visual defects, including deep neural networks, enable the processing of otherwise small defect patterns in pixel size on high-resolution imagery. The emergence of these often rarely occurring defect patterns explains the general need for labeled data corpora. To alleviate this issue and advance the current state of the art in unsupervised visual inspection, this work proposes a DifferNet-based solution enhanced with attention modules: AttentDifferNet. It improves image-level detection and classification capabilities on three visual anomaly detection datasets for industrial inspection: InsPLAD-fault, MVTec AD, and Semiconductor Wafer. In comparison to the state of the art, AttentDifferNet achieves improved results, which are, in turn, highlighted throughout our quali-quantitative study. Our quantitative evaluation shows an average improvement - compared to DifferNet - of 1.77 +- 0.25 percentage points in overall AUROC considering all three datasets, reaching SOTA results in InsPLAD-fault, an industrial inspection in-the-wild dataset. As our variants to AttentDifferNet show great prospects in the context of currently investigated approaches, a baseline is formulated, emphasizing the importance of attention for industrial anomaly detection both in the wild and in controlled environments.

The asymmetrical retrieval setting is a well suited solution for resource constrained applications such as face recognition and image retrieval. In this setting, a large model is used for indexing the gallery while a lightweight model is used for querying. The key principle in such systems is ensuring that both models share the same embedding space. Most methods in this domain are based on knowledge distillation. While useful, they suffer from several drawbacks: they are upper-bounded by the performance of the single best model found and cannot be extended to use an ensemble of models in a straightforward manner. In this paper we present an approach that does not rely on knowledge distillation, rather it utilizes embedding transformation models. This allows the use of N independently trained and diverse gallery models (e.g., trained on different datasets or having a different architecture) and a single query model. As a result, we improve the overall accuracy beyond that of any single model while maintaining a low computational budget for querying. Additionally, we propose a gallery image rejection method that utilizes the diversity between multiple transformed embeddings to estimate the uncertainty of gallery images.

Adversarial patches threaten visual AI models in the real world. The number of patches in a patch attack is variable and determines the attack's potency in a specific environment. Most existing defenses assume a single patch in the scene, and the multiple patch scenario are shown to overcome them. This paper presents a model-agnostic defense against patch attacks based on total variation for image resurfacing (TVR). The TVR is an image-cleansing method that processes images to remove probable adversarial regions. TVR can be utilized solely or augmented with a defended model, providing multi-level security for robust prediction. TVR nullifies the influence of patches in a single image scan with no prior assumption on the number of patches in the scene. We validate TVR on the ImageNet-Patch benchmark dataset and with real-world physical objects, demonstrating its ability to mitigate patch attack.

The robustness of image classifiers is essential to their deployment in the real world. The ability to assess this resilience to manipulations or deviations from the training data is thus crucial. These modifications have traditionally consisted of minimal changes that still manage to fool classifiers, and modern approaches are increasingly robust to them. Semantic manipulations that modify elements of an image in meaningful ways have thus gained traction for this purpose. However, they have primarily been limited to style, color, or attribute changes. While expressive, these manipulations do not make use of the full capabilities of a pretrained generative model. In this work, we aim to bridge this gap. We show how a pretrained image generator can be used to semantically manipulate images in a detailed, diverse, and photorealistic way while still preserving the class of the original image. Inspired by recent GAN-based image inversion methods, we propose a method called Adversarial Pivotal Tuning (APT). Given an image, APT first finds a pivot latent space input that reconstructs the image using a pretrained generator. It then adjusts the generator's weights to create small yet semantic manipulations in order to fool a pretrained classifier. APT preserves the full expressive editing capabilities of the generative model. We demonstrate that APT is capable of a wide range of class-preserving semantic image manipulations that fool a variety of pretrained classifiers. Finally, we show that classifiers that are robust to other benchmarks are not robust to APT manipulations and suggest a method to improve them.

Exploded views are powerful tools for visualizing the assembly and disassembly of complex objects, widely used in technical illustrations, assembly instructions, and product presentations. Previous methods for automating the creation of exploded views are either slow and computationally costly or compromise on accuracy. Therefore, the construction of exploded views is typically a manual process. In this paper, we propose a novel approach for automatically predicting the direction of parts in an exploded view using deep learning. To achieve this, we introduce a new dataset, AssemblyNet, which contains point cloud data sampled from 3D models of real-world assemblies, including water pumps, mixed industrial assemblies, and LEGO models. The AssemblyNet dataset includes a total of 44 assemblies, separated into 495 subassemblies with a total of 5420 parts. We provide ground truth labels for regression and classification, representing the directions in which the parts are moved in the exploded views. We also provide performance benchmarks using various state-of-the-art models for shape classification on point clouds and propose a novel two-path network architecture. Project page available at https://github.com/jgaarsdal/AssemblyNet

The task of Visual Question Answering (VQA) has been studied extensively on general-domain real-world images. Transferring insights from general domain VQA to the art domain (ArtVQA) is non-trivial, as the latter requires models to identify abstract concepts, details of brushstrokes and styles of paintings in the visual data as well as possess background knowledge about art. This is exacerbated by the lack of high-quality datasets. In this work, we shed light on hidden linguistic biases in the AQUA dataset, which is the only publicly available benchmark dataset for ArtVQA. As a result, the majority of questions can be answered without consulting the visual information, making the "V" in ArtVQA rather insignificant. In order to counter this problem, we create a simple, yet practical dataset, ArtQuest, using structured information from the SemArt collection. Our dataset and the pipeline to reproduce our results are publicly available at https://github.com/bletib/artquest.

Machine Learning (ML) models become vulnerable to Model Stealing Attacks (MSA) when they are deployed as a service. In such attacks, the deployed model is queried repeatedly to build a labelled dataset. This dataset allows the attacker to train a thief model that mimics the original model. To maximize query efficiency, the attacker has to select the most informative subset of data points from the pool of available data. Existing attack strategies utilize approaches like Active Learning and Semi-Supervised learning to minimize costs. However, in the black-box setting, these approaches may select sub-optimal samples as they train only one thief model. Depending on the thief model's capacity and the data it was pretrained on, the model might even select noisy samples that harm the learning process. In this work, we explore the usage of an ensemble of deep learning models as our thief model. We call our attack Army of Thieves(AOT) as we train multiple models with varying complexities to leverage the crowd's wisdom. Based on the ensemble's collective decision, uncertain samples are selected for querying, while the most confident samples are directly included in the training data. Our approach is the first one to utilize an ensemble of thief models to perform model extraction. We outperform the base approaches of existing state-of-the-art methods by at least 3% and achieve a 21% higher adversarial sample transferability than previous work for models trained on the CIFAR-10 dataset.

The deep learning field is converging towards the use of general foundation models that can be easily adapted for diverse tasks. While this paradigm shift has become common practice within the field of natural language processing, progress has been slower in computer vision. In this paper we attempt to address this issue by investigating the transferability of various state-of-the-art foundation models to medical image classification tasks. Specifically, we evaluate the performance of five foundation models, namely SAM, SEEM, DINOv2, BLIP, and OpenCLIP across four well-established medical imaging datasets. We explore different training settings to fully harness the potential of these models. Our study shows mixed results. DINOv2 consistently outperforms the standard practice of ImageNet pretraining. However, other foundation models failed to consistently beat this established baseline indicating limitations in their transferability to medical image classification tasks.

Cross-view matching techniques for image geolocalization attempt to match features in ground level imagery against a collection of satellite images to determine the position of given query image. We present a novel cross-view image matching approach called ArcGeo which introduces a batch-all angular margin loss and several train-time strategies including large-scale pretraining and FoV-based data augmentation. This allows our model to perform well even in challenging cases with limited field-of-view (FoV). Further, we evaluate multiple model architectures, data augmentation approaches and optimization strategies to train a deep cross-view matching network, specifically optimized for limited FoV cases. In low FoV experiments (FoV = 90deg) our method improves top-1 image recall rate on the CVUSA dataset from 30.12% to 43.08%. We also demonstrate improved performance over the state-of-the-art techniques for panoramic cross-view retrieval, improving top-1 recall from 95.43% to 96.06% on the CVUSA dataset and from 64.52% to 79.88% on the CVACT test dataset. Lastly, we evaluate the role of large-scale pretraining for improved robustness. With appropriate pretraining on external data, our model improves top-1 recall dramatically to 66.83% for FoV = 90deg test case on CVUSA, an increase of over twice what is reported by existing approaches.

Archaeology is an intriguing domain for computer vision. It suffers not only from shortage in (labeled) data, but also from highly-challenging data, which is often extremely abraded and damaged. This paper proposes a novel semi-supervised model for classification and retrieval of images of archaeological artifacts. This model utilizes unique data that exists in the domain--manual drawings made by special artists. These are used during training to implicitly transfer the domain knowledge from the drawings to their corresponding images, improving their classification results. We show that while learning how to classify, our model also learns how to generate drawings of the artifacts, an important documentation task, which is currently performed manually. Last but not least, we collected a new dataset of stamp-seals of the Southern Levant. Our code and dataset are publicly available.

Face super-resolution (FSR) is a critical technique for enhancing low-resolution facial images and has significant implications for face-related tasks. However, existing FSR methods are limited by fixed up-sampling scales and sensitivity to input size variations. To address these limitations, this paper introduces an Arbitrary-Resolution and Arbitrary-Scale FSR method with implicit representation networks (ARASFSR), featuring three novel designs. First, ARASFSR employs 2D deep features, local relative coordinates, and up-sampling scale ratios to predict RGB values for each target pixel, allowing super-resolution at any up-sampling scale. Second, a local frequency estimation module captures high-frequency facial texture information to reduce the spectral bias effect. Lastly, a global coordinate modulation module guides FSR to leverage prior knowledge of facial structure effectively and achieve resolution adaptation. Quantitative and qualitative evaluations demonstrate the robustness of ARASFSR over existing state-of-the-art methods while super-resolving facial images across various input sizes and up-sampling scales.

To date, the comparison of Statistical Shape Models (SSMs) is often solely performance-based, carried out by means of simplistic metrics such as compactness, generalization, or specificity. Any similarities or differences between the actual shape spaces can neither be visualized nor quantified. In this paper, we present a new method to qualitatively compare two linear SSMs in dense correspondence by computing approximate intersection spaces and set-theoretic differences between the (hyper-ellipsoidal) allowable shape domains spanned by the models. To this end, we approximate the distribution of shapes lying in the intersection space using Markov chain Monte Carlo and subsequently apply Principal Component Analysis (PCA) to the posterior samples, eventually yielding a new SSM of the intersection space. We estimate differences between linear SSMs in a similar manner; here, however, the resulting spaces are no longer convex and we do not apply PCA but instead use the posterior samples for visualization. We showcase the proposed algorithm qualitatively by computing and analyzing intersection spaces and differences between publicly available face models, focusing on gender-specific male and female as well as identity and expression models. Our quantitative evaluation based on SSMs built from synthetic and real-world data sets provides detailed evidence that the introduced method is able to recover ground-truth intersection spaces and differences accurately.