International Journal of Computer Vision最新文献

Deepfake techniques pose a significant threat to personal privacy and social security. To mitigate these risks, various defensive techniques have been introduced, including passive methods through fake detection and proactive methods through adding invisible perturbations. Recent proactive methods mainly focus on face manipulation but perform poorly against face swapping, as face swapping involves the more complex process of identity information transfer. To address this issue, we develop a novel privacy-preserving framework, named Anti-Fake Vaccine, to protect the facial images against the malicious face swapping. This new proactive technique dynamically fuses visual corruption and content misdirection, significantly enhancing protection performance. Specifically, we first formulate constraints from two distinct perspectives: visual quality and identity semantics. The visual perceptual constraint targets image quality degradation in the visual space, while the identity similarity constraint induces erroneous alterations in the semantic space. We then introduce a multi-objective optimization solution to effectively balance the allocation of adversarial perturbations generated according to these constraints. To further improving performance, we develop an additive perturbation strategy to discover the shared adversarial perturbations across diverse face swapping models. Extensive experiments on the CelebA-HQ and FFHQ datasets demonstrate that our method exhibits superior generalization capabilities across diverse face swapping models, including commercial ones.

3D biometric techniques on finger traits have become a new trend and have demonstrated a powerful ability for recognition and anti-counterfeiting. Existing methods follow an explicit 3D pipeline that reconstructs the models first and then extracts features from 3D models. However, these explicit 3D methods suffer from the following problems: 1) Inevitable information dropping during 3D reconstruction; 2) Tight coupling between specific hardware and algorithm for 3D reconstruction. It leads us to a question: Is it indispensable to reconstruct 3D information explicitly in recognition tasks? Hence, we consider this problem in an implicit manner, leaving the nerve-wracking 3D reconstruction problem for learnable neural networks with the help of neural radiance fields (NeRFs). We propose FingerNeRF, a novel generalizable NeRF for 3D finger biometrics. To handle the shape-radiance ambiguity problem that may result in incorrect 3D geometry, we aim to involve extra geometric priors based on the correspondence of binary finger traits like fingerprints or finger veins. First, we propose a novel Trait Guided Transformer (TGT) module to enhance the feature correspondence with the guidance of finger traits. Second, we involve extra geometric constraints on the volume rendering loss with the proposed Depth Distillation Loss and Trait Guided Rendering Loss. To evaluate the performance of the proposed method on different modalities, we collect two new datasets: SCUT-Finger-3D with finger images and SCUT-FingerVein-3D with finger vein images. Moreover, we also utilize the UNSW-3D dataset with fingerprint images for evaluation. In experiments, our FingerNeRF can achieve 4.37% EER on SCUT-Finger-3D dataset, 8.12% EER on SCUT-FingerVein-3D dataset, and 2.90% EER on UNSW-3D dataset, showing the superiority of the proposed implicit method in 3D finger biometrics.

This paper introduces a new framework for surface analysis derived from the general setting of elastic Riemannian metrics on shape spaces. Traditionally, those metrics are defined over the infinite dimensional manifold of immersed surfaces and satisfy specific invariance properties enabling the comparison of surfaces modulo shape preserving transformations such as reparametrizations. The specificity of our approach is to restrict the space of allowable transformations to predefined finite dimensional bases of deformation fields. These are estimated in a data-driven way so as to emulate specific types of surface transformations. This allows us to simplify the representation of the corresponding shape space to a finite dimensional latent space. However, in sharp contrast with methods involving e.g. mesh autoencoders, the latent space is equipped with a non-Euclidean Riemannian metric inherited from the family of elastic metrics. We demonstrate how this model can be effectively implemented to perform a variety of tasks on surface meshes which, importantly, does not assume these to be pre-registered or to even have a consistent mesh structure. We specifically validate our approach on human body shape and pose data as well as human face and hand scans for problems such as shape registration, interpolation, motion transfer or random pose generation.

As a crucial topic of high-level video understanding, weakly supervised online activity detection (WS-OAD) involves identifying the ongoing behaviors moment-to-moment in streaming videos, trained with solely cheap video-level annotations. It is essentially a challenging task, which requires addressing the entangled issues of the weakly supervised settings and online constraints. In this paper, we tackle the WS-OAD task from the knowledge-distillation (KD) perspective, which trains an online student detector to distill dual-level knowledge from a weakly supervised offline teacher model. To guarantee the completeness of knowledge transfer, we improve the vanilla KD framework from two aspects. First, we introduce an external memory bank to maintain the long-term activity prototypes, which serves as a bridge to align the activity semantics learned from the offline teacher and online student models. Second, to compensate the missing contexts of unseen near future, we leverage a curriculum learning paradigm to gradually train the online student detector to anticipate the future activity semantics. By dynamically scheduling the provided auxiliary future states, the online detector progressively distills contextual information from the offline model in an easy-to-hard course. Extensive experimental results on three public data sets demonstrate the superiority of our proposed method over the competing methods.

Multimodal based person re-identification (ReID) has garnered increasing attention in recent years. However, the integration of visual and textual information encounters significant challenges. Biases in feature integration are frequently observed in existing methods, resulting in suboptimal performance and restricted generalization across a spectrum of ReID tasks. At the same time, since there is a domain gap between the datasets used by the pretraining model and the ReID datasets, it has a certain impact on the performance. In response to these challenges, we proposed a dynamic attention vision-language transformer network for the ReID task. In this network, a novel image-text dynamic attention module (ITDA) is designed to promote unbiased feature integration by dynamically assigning the importance of image and text representations. Additionally, an adapter module is adopted to address the domain gap between pretraining datasets and ReID datasets. Our network can capture complex connections between visual and textual information and achieve satisfactory performance. We conducted numerous experiments on ReID benchmarks to demonstrate the efficacy of our proposed method. The experimental results show that our method achieves state-of-the-art performance, surpassing existing integration strategies. These findings underscore the critical role of unbiased feature dynamic integration in enhancing the capabilities of multimodal based ReID models.

This work focuses on generating high-quality images with specific style of reference images and content of provided textual descriptions. Current leading algorithms, i.e., DreamBooth and LoRA, require fine-tuning for each style, leading to time-consuming and computationally expensive processes. In this work, we propose StyleAdapter, a unified stylized image generation model capable of producing a variety of stylized images that match both the content of a given prompt and the style of reference images, without the need for per-style fine-tuning. It introduces a two-path cross-attention (TPCA) module to separately process style information and textual prompt, which cooperate with a semantic suppressing vision model (SSVM) to suppress the semantic content of style images. In this way, it can ensure that the prompt maintains control over the content of the generated images, while also mitigating the negative impact of semantic information in style references. This results in the content of the generated image adhering to the prompt, and its style aligning with the style references. Besides, our StyleAdapter can be integrated with existing controllable synthesis methods, such as T2I-adapter and ControlNet, to attain a more controllable and stable generation process. Extensive experiments demonstrate the superiority of our method over previous works.

Face recognition has witnessed remarkable advancements in recent years, thanks to the development of deep learning techniques. However, an off-the-shelf face recognition model as a commercial service could be stolen by model stealing attacks, posing great threats to the rights of the model owner. Model fingerprinting, as a model stealing detection method, aims to verify whether a suspect model is stolen from the victim model, gaining more and more attention nowadays. Previous methods always utilize transferable adversarial examples as the model fingerprint, but this method is known to be sensitive to adversarial defense and transfer learning techniques. To address this issue, we consider the pairwise relationship between samples instead and propose a novel yet simple model stealing detection method based on SAmple Correlation (SAC). Specifically, we present SAC-JC that selects JPEG compressed samples as model inputs and calculates the correlation matrix among their model outputs. Extensive results validate that SAC successfully defends against various model stealing attacks in deep face recognition, encompassing face verification and face emotion recognition, exhibiting the highest performance in terms of AUC, p-value and F1 score. Furthermore, we extend our evaluation of SAC-JC to object recognition datasets including Tiny-ImageNet and CIFAR10, which also demonstrates the superior performance of SAC-JC to previous methods. The code will be available at https://github.com/guanjiyang/SAC_JC.

Significant advancements have been achieved in the realm of large-scale pre-trained text-to-video Diffusion Models (VDMs). However, previous methods either rely solely on pixel-based VDMs, which come with high computational costs, or on latent-based VDMs, which often struggle with precise text-video alignment. In this paper, we are the first to propose a hybrid model, dubbed as Show-1, which marries pixel-based and latent-based VDMs for text-to-video generation. Our model first uses pixel-based VDMs to produce a low-resolution video of strong text-video correlation. After that, we propose a novel expert translation method that employs the latent-based VDMs to further upsample the low-resolution video to high resolution, which can also remove potential artifacts and corruptions from low-resolution videos. Compared to latent VDMs, Show-1 can produce high-quality videos of precise text-video alignment; Compared to pixel VDMs, Show-1 is much more efficient (GPU memory usage during inference is 15 G vs. 72 G). Furthermore, our Show-1 model can be readily adapted for motion customization and video stylization applications through simple temporal attention layer finetuning. Our model achieves state-of-the-art performance on standard video generation benchmarks. Code of Show-1 is publicly available and more videos can be found here.

Neural implicit fields have recently shown increasing success in representing, learning and analysis of 3D shapes. Signed distance fields and occupancy fields are still the preferred choice of implicit representations with well-studied properties, despite their restriction to closed surfaces. With neural networks, unsigned distance fields as well as several other variations and training principles have been proposed with the goal to represent all classes of shapes. In this paper, we develop a novel and yet a fundamental representation considering unit vectors in 3D space and call it Vector Field (VF). At each point in (mathbb {R}^3), VF is directed to the closest point on the surface. We theoretically demonstrate that VF can be easily transformed to surface density by computing the flux density. Unlike other standard representations, VF directly encodes an important physical property of the surface, its normal. We further show the advantages of VF representation, in learning open, closed, or multi-layered surfaces. We show that, thanks to the continuity property of the neural optimization with VF, a separate distance field becomes unnecessary for extracting surfaces from the implicit field via Marching Cubes. We compare our method on several datasets including ShapeNet where the proposed new neural implicit field shows superior accuracy in representing any type of shape, outperforming other standard methods. Codes are available at https://github.com/edomel/ImplicitVF.

We describe a protocol to study text-to-video retrieval training with unlabeled videos, where we assume (i) no access to labels for any videos, i.e., no access to the set of ground-truth captions, but (ii) access to labeled images in the form of text. Using image expert models is a realistic scenario given that annotating images is cheaper therefore scalable, in contrast to expensive video labeling schemes. Recently, zero-shot image experts such as CLIP have established a new strong baseline for video understanding tasks. In this paper, we make use of this progress and instantiate the image experts from two types of models: a text-to-image retrieval model to provide an initial backbone, and image captioning models to provide supervision signal into unlabeled videos. We show that automatically labeling video frames with image captioning allows text-to-video retrieval training. This process adapts the features to the target domain at no manual annotation cost, consequently outperforming the strong zero-shot CLIP baseline. During training, we sample captions from multiple video frames that best match the visual content, and perform a temporal pooling over frame representations by scoring frames according to their relevance to each caption. We conduct extensive ablations to provide insights and demonstrate the effectiveness of this simple framework by outperforming the CLIP zero-shot baselines on text-to-video retrieval on three standard datasets, namely ActivityNet, MSR-VTT, and MSVD. Code and models will be made publicly available.