Boyi Jiang

Zile Wang, Zexiang Liu, Jaixing Li et al.

With the advancement of interactive video generation, diffusion models have increasingly demonstrated their potential as world models. However, existing approaches still struggle to simultaneously achieve memory-enabled long-term temporal consistency and high-resolution real-time generation, limiting their applicability in real-world scenarios. To address this, we present Matrix-Game 3.0, a memory-augmented interactive world model designed for 720p real-time longform video generation. Building upon Matrix-Game 2.0, we introduce systematic improvements across data, model, and inference. First, we develop an upgraded industrial-scale infinite data engine that integrates Unreal Engine-based synthetic data, large-scale automated collection from AAA games, and real-world video augmentation to produce high-quality Video-Pose-Action-Prompt quadruplet data at scale. Second, we propose a training framework for long-horizon consistency: by modeling prediction residuals and re-injecting imperfect generated frames during training, the base model learns self-correction; meanwhile, camera-aware memory retrieval and injection enable the base model to achieve long horizon spatiotemporal consistency. Third, we design a multi-segment autoregressive distillation strategy based on Distribution Matching Distillation (DMD), combined with model quantization and VAE decoder pruning, to achieve efficient real-time inference. Experimental results show that Matrix-Game 3.0 achieves up to 40 FPS real-time generation at 720p resolution with a 5B model, while maintaining stable memory consistency over minute-long sequences. Scaling up to a 2x14B model further improves generation quality, dynamics, and generalization. Our approach provides a practical pathway toward industrial-scale deployable world models.

Hongyang Wei, Baixin Xu, Hongbo Liu et al.

Recent advances in multimodal models have demonstrated impressive capabilities in unified image generation and editing. However, many prominent open-source models prioritize scaling model parameters over optimizing training strategies, limiting their efficiency and performance. In this work, we present UniPic2-SD3.5M-Kontext, a 2B-parameter DiT model based on SD3.5-Medium, which achieves state-of-the-art image generation and editing while extending seamlessly into a unified multimodal framework. Our approach begins with architectural modifications to SD3.5-Medium and large-scale pre-training on high-quality data, enabling joint text-to-image generation and editing capabilities. To enhance instruction following and editing consistency, we propose a novel Progressive Dual-Task Reinforcement strategy (PDTR), which effectively strengthens both tasks in a staged manner. We empirically validate that the reinforcement phases for different tasks are mutually beneficial and do not induce negative interference. After pre-training and reinforcement strategies, UniPic2-SD3.5M-Kontext demonstrates stronger image generation and editing capabilities than models with significantly larger generation parameters-including BAGEL (7B) and Flux-Kontext (12B). Furthermore, following the MetaQuery, we connect the UniPic2-SD3.5M-Kontext and Qwen2.5-VL-7B via a connector and perform joint training to launch a unified multimodal model UniPic2-Metaquery. UniPic2-Metaquery integrates understanding, generation, and editing, achieving top-tier performance across diverse tasks with a simple and scalable training paradigm. This consistently validates the effectiveness and generalizability of our proposed training paradigm, which we formalize as Skywork UniPic 2.0.

Boyi Jiang, Yang Hong, Hujun Bao et al.

We propose SelfRecon, a clothed human body reconstruction method that combines implicit and explicit representations to recover space-time coherent geometries from a monocular self-rotating human video. Explicit methods require a predefined template mesh for a given sequence, while the template is hard to acquire for a specific subject. Meanwhile, the fixed topology limits the reconstruction accuracy and clothing types. Implicit representation supports arbitrary topology and can represent high-fidelity geometry shapes due to its continuous nature. However, it is difficult to integrate multi-frame information to produce a consistent registration sequence for downstream applications. We propose to combine the advantages of both representations. We utilize differential mask loss of the explicit mesh to obtain the coherent overall shape, while the details on the implicit surface are refined with the differentiable neural rendering. Meanwhile, the explicit mesh is updated periodically to adjust its topology changes, and a consistency loss is designed to match both representations. Compared with existing methods, SelfRecon can produce high-fidelity surfaces for arbitrary clothed humans with self-supervised optimization. Extensive experimental results demonstrate its effectiveness on real captured monocular videos. The source code is available at https://github.com/jby1993/SelfReconCode.

Boyi Jiang, Juyong Zhang, Yang Hong et al.

In this paper, we consider the problem to automatically reconstruct garment and body shapes from a single near-front view RGB image. To this end, we propose a layered garment representation on top of SMPL and novelly make the skinning weight of garment independent of the body mesh, which significantly improves the expression ability of our garment model. Compared with existing methods, our method can support more garment categories and recover more accurate geometry. To train our model, we construct two large scale datasets with ground truth body and garment geometries as well as paired color images. Compared with single mesh or non-parametric representation, our method can achieve more flexible control with separate meshes, makes applications like re-pose, garment transfer, and garment texture mapping possible. Code and some data is available at https://github.com/jby1993/BCNet.

Yang Hong, Juyong Zhang, Boyi Jiang et al.

In this paper, we propose StereoPIFu, which integrates the geometric constraints of stereo vision with implicit function representation of PIFu, to recover the 3D shape of the clothed human from a pair of low-cost rectified images. First, we introduce the effective voxel-aligned features from a stereo vision-based network to enable depth-aware reconstruction. Moreover, the novel relative z-offset is employed to associate predicted high-fidelity human depth and occupancy inference, which helps restore fine-level surface details. Second, a network structure that fully utilizes the geometry information from the stereo images is designed to improve the human body reconstruction quality. Consequently, our StereoPIFu can naturally infer the human body's spatial location in camera space and maintain the correct relative position of different parts of the human body, which enables our method to capture human performance. Compared with previous works, our StereoPIFu significantly improves the robustness, completeness, and accuracy of the clothed human reconstruction, which is demonstrated by extensive experimental results.

Boyi Jiang, Juyong Zhang, Jianfei Cai et al.

Human bodies exhibit various shapes for different identities or poses, but the body shape has certain similarities in structure and thus can be embedded in a low-dimensional space. This paper presents an autoencoder-like network architecture to learn disentangled shape and pose embedding specifically for the 3D human body. This is inspired by recent progress of deformation-based latent representation learning. To improve the reconstruction accuracy, we propose a hierarchical reconstruction pipeline for the disentangling process and construct a large dataset of human body models with consistent connectivity for the learning of the neural network. Our learned embedding can not only achieve superior reconstruction accuracy but also provide great flexibility in 3D human body generation via interpolation, bilinear interpolation, and latent space sampling. The results from extensive experiments demonstrate the powerfulness of our learned 3D human body embedding in various applications.

Boyi Jiang, Juyong Zhang, Bailin Deng et al.

In this paper, we present a deep learning based image feature extraction method designed specifically for face images. To train the feature extraction model, we construct a large scale photo-realistic face image dataset with ground-truth correspondence between multi-view face images, which are synthesized from real photographs via an inverse rendering procedure. The deep face feature (DFF) is trained using correspondence between face images rendered from different views. Using the trained DFF model, we can extract a feature vector for each pixel of a face image, which distinguishes different facial regions and is shown to be more effective than general-purpose feature descriptors for face-related tasks such as matching and alignment. Based on the DFF, we develop a robust face alignment method, which iteratively updates landmarks, pose and 3D shape. Extensive experiments demonstrate that our method can achieve state-of-the-art results for face alignment under highly unconstrained face images.

Yudong Guo, Juyong Zhang, Jianfei Cai et al.

With the powerfulness of convolution neural networks (CNN), CNN based face reconstruction has recently shown promising performance in reconstructing detailed face shape from 2D face images. The success of CNN-based methods relies on a large number of labeled data. The state-of-the-art synthesizes such data using a coarse morphable face model, which however has difficulty to generate detailed photo-realistic images of faces (with wrinkles). This paper presents a novel face data generation method. Specifically, we render a large number of photo-realistic face images with different attributes based on inverse rendering. Furthermore, we construct a fine-detailed face image dataset by transferring different scales of details from one image to another. We also construct a large number of video-type adjacent frame pairs by simulating the distribution of real video data. With these nicely constructed datasets, we propose a coarse-to-fine learning framework consisting of three convolutional networks. The networks are trained for real-time detailed 3D face reconstruction from monocular video as well as from a single image. Extensive experimental results demonstrate that our framework can produce high-quality reconstruction but with much less computation time compared to the state-of-the-art. Moreover, our method is robust to pose, expression and lighting due to the diversity of data.