Despite the increasing popularity of avatar systems such as Snapchat Bitmojis, existing production avatar platforms face several limitations, such as a limited number of predefined assets, tedious customization processes, and inefficient rendering requirements. Addressing these shortcomings, we introduce Snapmoji, an avatar generation system that instantly creates 3D avatars, and enables customization in a process we call dual-stylization. Snapmoji first maps a selfie of a user to a primary avatar (e.g., Bitmoji style) using a new technique we name Gaussian Domain Adaptation (GDA), then applies a secondary style (e.g., skeleton, yarn, toy) to the primary avatar, all while preserving the user’s identity. The generated 3D avatars can then be rendered an animated on mobile devices at 30–40 FPS.

This paper describes a Reliable Learning Framework (RLF) for the 1st Multimodal Emotion and Intent Joint Understanding (MEIJU) Challenge at ICASSP 2025. Our proposed RLF includes a Hierarchical Interaction Network and a Reliable Fusion Strategy. The former can excavate emotion and intent cues from the high-level semantic features of multimodal data (video, audio, and text) generated by pretrained Large Language Models (LMMs), to enhance their representations, and the latter reliably integrates multiple predictions to further improve the robustness of emotion and intent understanding. Our RLF method achieved first place on Track 2 (Mandarin) of MEIJU, with performance scores for emotion, intent, and joint recognition reaching 0.7285, 0.7456, and 0.7370.

We present T2Bs, a framework for generating high-quality, animatable character head morphable models from text by combining static text-to-3D generation with video diffusion. Text-to-3D models produce detailed static geometry but lack motion synthesis, while video diffusion models generate motion with temporal and multi-view geometric inconsistencies. T2Bs bridges this gap by leveraging deformable 3D Gaussian splatting to align static 3D assets with video outputs. By constraining motion with static geometry and employing a view-dependent deformation MLP, T2Bs (i) outperforms existing 4D generation methods in accuracy and expressiveness while reducing video artifacts and view inconsistencies, and (ii) reconstructs smooth, coherent, fully registered 3D geometries designed to scale for building morphable models with diverse, realistic facial motions. This enables synthesizing expressive, animatable character heads that surpass current 4D generation techniques.

We present SplatFace, a novel Gaussian splatting framework designed for 3D human face reconstruction without reliance on accurate pre-determined geometry. Our method is designed to simultaneously deliver both high-quality novel view rendering and accurate 3D mesh reconstructions. We incorporate a generic 3D Morphable Model (3DMM) to provide a surface geometric structure, making it possible to reconstruct faces with a limited set of input images. We introduce a joint optimization strategy that refines both the Gaussians and the morphable surface through a synergistic non-rigid alignment process. A novel distance metric, splat-to-surface, is proposed to improve alignment by considering both the Gaussian position and covariance. The surface information is also utilized to incorporate a world-space densification process, resulting in superior reconstruction quality. Our experimental analysis demonstrates that the proposed method is competitive with both other Gaussian splatting techniques in novel view synthesis and other 3D reconstruction methods in producing 3D face meshes with high geometric precision.

Face models are widely used in image processing and other domains. The input data to create a 3D face model ranges from accurate laser scans to simple 2D RGB photographs. These input data types are typically deficient either due to missing regions, or because they are underconstrained. As a result, reconstruction methods include embedded priors encoding the valid domain of faces. System designers must choose a source of input data and then choose a reconstruction method to obtain a usable 3D face. If a particular application domain requires accuracy X, which kinds of input data are suitable? Does the input data need to be 3D, or will 2D data suffice? This paper takes a step toward answering these questions using synthetic data. A ground truth dataset is used to analyze accuracy obtainable from 2D landmarks, 3D landmarks, low quality 3D, high quality 3D, texture color, normals, dense 2D image data, and when regions of the face are missing. Since the data is synthetic it can be analyzed both with and without measurement error. This idealized synthetic analysis is then compared to real results from several methods for constructing 3D faces from 2D photographs. The experimental results suggest that accuracy is severely limited when only 2D raw input data exists.

The highest quality 3D face reconstructions are produced using multi-view stereo methods, reporting errors below 0.5mm. Unfortunately, these methods typically employ dozens of high-resolution cameras in a large laboratory capture gantry. In contrast, monocular 3D face reconstruction using sophisticated deep learning models are suited for casual mobile phone imaging outside the lab and report a mean error of 1-2mm.This paper investigates whether classic stereo methods can be used in scenarios with only a few low-resolution images available. We expect to find that it cannot since multi-view stereo performs well only when many high-resolution images are provided. When only two low-resolution images are available, stereo produces very noisy results which are not directly usable. Surprisingly, however, our analysis shows that this visually noisy data has lower error than comparison state-of-the-art methods. We find that the visual artifacts from stereo can be removed using a morphable face model to constrain face shape.

The capability to generate clear and bright images in low light situations is crucial for photographers, engineers, and researchers. When it is not possible to modify the imaging conditions, an algorithm to enhance images is needed. Traditional methods require manually adjusting parameters to tune the image. Supervised learning methods need to collect a large amount of paired data for training. In this paper, we demonstrate an semi-supervised method for low light image enhancement, using a chain of cycle consistent generators. We show the effectiveness of our method by comparing it to existing image enhancement methods, both using standard image quality metrics and by using human perceptual judgements. We include an ablation study for features in our model. Our proposed method is computationally efficient and does not require paired training data.

Face models are widely used in image processing and other domains. The input data to create a 3D face model ranges from accurate laser scans to simple 2D RGB photographs. System designers must choose a source of input data and then choose a reconstruction method to obtain a usable 3D face. If a particular application domain requires accuracy X, which kinds of input data are suitable? This paper takes a step toward answering this question. A variety of common input data types such as 2D landmarks and 3D scans are constructed from an existing high quality dataset. A morphable face model is then used to reconstruct 3D faces. By comparing to ground truth, an analysis of the relative error between different data types is obtained.