GAF: Gaussian Avatar Reconstruction from Monocular Videos via Multi-view Diffusion

We propose a novel approach for reconstructing animatable 3D Gaussian avatars from monocular videos captured by commodity devices like smartphones. Photorealistic 3D head avatar reconstruction from such recordings is challenging due to limited observations, which leaves unobserved regions under-constrained and can lead to artifacts in novel views. To address this problem, we introduce a multi-view head diffusion model, leveraging its priors to fill in missing regions and ensure view consistency in Gaussian splatting renderings. To enable precise viewpoint control, we use normal maps rendered from FLAME-based head reconstruction, which provides pixel-aligned inductive biases. We also condition the diffusion model on VAE features extracted from the input image to preserve details of facial identity and appearance. For Gaussian avatar reconstruction, we distill multi-view diffusion priors by using iteratively denoised images as pseudo-ground truths, effectively mitigating over-saturation issues. To further improve photorealism, we apply latent upsampling to refine the denoised latent before decoding it into an image. We evaluate our method on the NeRSemble dataset, showing that GAF outperforms the previous state-of-the-art methods in novel view synthesis by a 5.34\% higher SSIM score. Furthermore, we demonstrate higher-fidelity avatar reconstructions from monocular videos captured on commodity devices.

HeadCraft: Modeling High-Detail Shape Variations for Animated 3DMMs

Current advances in human head modeling allow to generate plausible-looking 3D head models via neural representations. Nevertheless, constructing complete high-fidelity head models with explicitly controlled animation remains an issue. Furthermore, completing the head geometry based on a partial observation, e.g. coming from a depth sensor, while preserving details is often problematic for the existing methods. We introduce a generative model for detailed 3D head meshes on top of an articulated 3DMM which allows explicit animation and high-detail preservation at the same time. Our method is trained in two stages. First, we register a parametric head model with vertex displacements to each mesh of the recently introduced NPHM dataset of accurate 3D head scans. The estimated displacements are baked into a hand-crafted UV layout. Second, we train a StyleGAN model in order to generalize over the UV maps of displacements. The decomposition of the parametric model and high-quality vertex displacements allows us to animate the model and modify it semantically. We demonstrate the results of unconditional generation and fitting to the full or partial observation. The project page is available at https://seva100.github.io/headcraft.

DPHMs: Diffusion Parametric Head Models for Depth-based Tracking

We introduce Diffusion Parametric Head Models (DPHMs), a generative model that enables robust volumetric head reconstruction and tracking from monocular depth sequences. While recent volumetric head models, such as NPHMs, can now excel in representing high-fidelity head geometries, tracking and reconstruction heads from real-world single-view depth sequences remains very challenging, as the fitting to partial and noisy observations is underconstrained. To tackle these challenges, we propose a latent diffusion-based prior to regularize volumetric head reconstruction and tracking. This prior-based regularizer effectively constrains the identity and expression codes to lie on the underlying latent manifold which represents plausible head shapes. To evaluate the effectiveness of the diffusion-based prior, we collect a dataset of monocular Kinect sequences consisting of various complex facial expression motions and rapid transitions. We compare our method to state-of-the-art tracking methods, and demonstrate improved head identity reconstruction as well as robust expression tracking.

Generating Context-Aware Natural Answers for Questions in 3D Scenes

3D question answering is a young field in 3D vision-language that is yet to be explored. Previous methods are limited to a pre-defined answer space and cannot generate answers naturally. In this work, we pivot the question answering task to a sequence generation task to generate free-form natural answers for questions in 3D scenes (Gen3DQA). To this end, we optimize our model directly on the language rewards to secure the global sentence semantics. Here, we also adapt a pragmatic language understanding reward to further improve the sentence quality. Our method sets a new SOTA on the ScanQA benchmark (CIDEr score 72.22/66.57 on the test sets).

3DShape2VecSet: A 3D Shape Representation for Neural Fields and Generative Diffusion Models

We introduce 3DShape2VecSet, a novel shape representation for neural fields designed for generative diffusion models. Our shape representation can encode 3D shapes given as surface models or point clouds, and represents them as neural fields. The concept of neural fields has previously been combined with a global latent vector, a regular grid of latent vectors, or an irregular grid of latent vectors. Our new representation encodes neural fields on top of a set of vectors. We draw from multiple concepts, such as the radial basis function representation and the cross attention and self-attention function, to design a learnable representation that is especially suitable for processing with transformers. Our results show improved performance in 3D shape encoding and 3D shape generative modeling tasks. We demonstrate a wide variety of generative applications: unconditioned generation, category-conditioned generation, text-conditioned generation, point-cloud completion, and image-conditioned generation.

On the Exploitation of Deepfake Model Recognition

Despite recent advances in Generative Adversarial Networks (GANs), with special focus to the Deepfake phenomenon there is no a clear understanding neither in terms of explainability nor of recognition of the involved models. In particular, the recognition of a specific GAN model that generated the deepfake image compared to many other possible models created by the same generative architecture (e.g. StyleGAN) is a task not yet completely addressed in the state-of-the-art. In this work, a robust processing pipeline to evaluate the possibility to point-out analytic fingerprints for Deepfake model recognition is presented. After exploiting the latent space of 50 slightly different models through an in-depth analysis on the generated images, a proper encoder was trained to discriminate among these models obtaining a classification accuracy of over 96%. Once demonstrated the possibility to discriminate extremely similar images, a dedicated metric exploiting the insights discovered in the latent space was introduced. By achieving a final accuracy of more than 94% for the Model Recognition task on images generated by models not employed in the training phase, this study takes an important step in countering the Deepfake phenomenon introducing a sort of signature in some sense similar to those employed in the multimedia forensics field (e.g. for camera source identification task, image ballistics task, etc).

TAFIM: Targeted Adversarial Attacks against Facial Image Manipulations

Face image manipulation methods, despite having many beneficial applications in computer graphics, can also raise concerns by affecting an individual's privacy or spreading disinformation. In this work, we propose a proactive defense to prevent face manipulation from happening in the first place. To this end, we introduce a novel data-driven approach that produces image-specific perturbations which are embedded in the original images. The key idea is that these protected images prevent face manipulation by causing the manipulation model to produce a predefined manipulation target (uniformly colored output image in our case) instead of the actual manipulation. Compared to traditional adversarial attacks that optimize noise patterns for each image individually, our generalized model only needs a single forward pass, thus running orders of magnitude faster and allowing for easy integration in image processing stacks, even on resource-constrained devices like smartphones. In addition, we propose to leverage a differentiable compression approximation, hence making generated perturbations robust to common image compression. We further show that a generated perturbation can simultaneously prevent against multiple manipulation methods.

Advances in Neural Rendering

Synthesizing photo-realistic images and videos is at the heart of computer graphics and has been the focus of decades of research. Traditionally, synthetic images of a scene are generated using rendering algorithms such as rasterization or ray tracing, which take specifically defined representations of geometry and material properties as input. Collectively, these inputs define the actual scene and what is rendered, and are referred to as the scene representation (where a scene consists of one or more objects). Example scene representations are triangle meshes with accompanied textures (e.g., created by an artist), point clouds (e.g., from a depth sensor), volumetric grids (e.g., from a CT scan), or implicit surface functions (e.g., truncated signed distance fields). The reconstruction of such a scene representation from observations using differentiable rendering losses is known as inverse graphics or inverse rendering. Neural rendering is closely related, and combines ideas from classical computer graphics and machine learning to create algorithms for synthesizing images from real-world observations. Neural rendering is a leap forward towards the goal of synthesizing photo-realistic image and video content. In recent years, we have seen immense progress in this field through hundreds of publications that show different ways to inject learnable components into the rendering pipeline. This state-of-the-art report on advances in neural rendering focuses on methods that combine classical rendering principles with learned 3D scene representations, often now referred to as neural scene representations. A key advantage of these methods is that they are 3D-consistent by design, enabling applications such as novel viewpoint synthesis of a captured scene. In addition to methods that handle static scenes, we cover neural scene representations for modeling non-rigidly deforming objects...

CAD-Deform: Deformable Fitting of CAD Models to 3D Scans

Shape retrieval and alignment are a promising avenue towards turning 3D scans into lightweight CAD representations that can be used for content creation such as mobile or AR/VR gaming scenarios. Unfortunately, CAD model retrieval is limited by the availability of models in standard 3D shape collections (e.g., ShapeNet). In this work, we address this shortcoming by introducing CAD-Deform, a method which obtains more accurate CAD-to-scan fits by non-rigidly deforming retrieved CAD models. Our key contribution is a new non-rigid deformation model incorporating smooth transformations and preservation of sharp features, that simultaneously achieves very tight fits from CAD models to the 3D scan and maintains the clean, high-quality surface properties of hand-modeled CAD objects. A series of thorough experiments demonstrate that our method achieves significantly tighter scan-to-CAD fits, allowing a more accurate digital replica of the scanned real-world environment while preserving important geometric features present in synthetic CAD environments.

Explicitly Modeled Attention Maps for Image Classification

Self-attention networks have shown remarkable progress in computer vision tasks such as image classification. The main benefit of the self-attention mechanism is the ability to capture long-range feature interactions in attention-maps. However, the computation of attention-maps requires a learnable key, query, and positional encoding, whose usage is often not intuitive and computationally expensive. To mitigate this problem, we propose a novel self-attention module with explicitly modeled attention-maps using only a single learnable parameter for low computational overhead. The design of explicitly modeled attention-maps using geometric prior is based on the observation that the spatial context for a given pixel within an image is mostly dominated by its neighbors, while more distant pixels have a minor contribution. Concretely, the attention-maps are parametrized via simple functions (e.g., Gaussian kernel) with a learnable radius, which is modeled independently of the input content. Our evaluation shows that our method achieves an accuracy improvement of up to 2.2% over the ResNet-baselines in ImageNet ILSVRC and outperforms other self-attention methods such as AA-ResNet152 (Bello et al., 2019) in accuracy by 0.9% with 6.4% fewer parameters and 6.7% fewer GFLOPs.