Creating controllable, photorealistic, and geometrically detailed digital doubles of real humans solely from video data is a key challenge in Computer Graphics and Vision, especially when real-time performance is required. Recent methods attach a neural radiance field (NeRF) to an articulated structure, e.g., a body model or a skeleton, to map points into a pose canonical space while conditioning the NeRF on the skeletal pose. These approaches typically parameterize the neural field with a multi-layer perceptron (MLP) leading to a slow runtime. To address this drawback, we propose TriHuman a novel human-tailored, deformable, and efficient tri-plane representation, which achieves real-time performance, state-of-the-art pose-controllable geometry synthesis as well as photorealistic rendering quality. At the core, we non-rigidly warp global ray samples into our undeformed tri-plane texture space, which effectively addresses the problem of global points being mapped to the same tri-plane locations. We then show how such a tri-plane feature representation can be conditioned on the skeletal motion to account for dynamic appearance and geometry changes. Our results demonstrate a clear step towards higher quality in terms of geometry and appearance modeling of humans as well as runtime performance.
相關內容
Performance：International Symposium on Computer Performance Modeling, Measurements and Evaluation。
Explanation：計算機性能建(jian)模、測量和(he)評估國際研討(tao)會。
Publisher：ACM。
SIT：
The share of online video traffic in global carbon dioxide emissions is growing steadily. To comply with the demand for video media, dedicated compression techniques are continuously optimized, but at the expense of increasingly higher computational demands and thus rising energy consumption at the video encoder side. In order to find the best trade-off between compression and energy consumption, modeling encoding energy for a wide range of encoding parameters is crucial. We propose an encoding time and energy model for SVT-AV1 based on empirical relations between the encoding time and video parameters as well as encoder configurations. Furthermore, we model the influence of video content by established content descriptors such as spatial and temporal information. We then use the predicted encoding time to estimate the required energy demand and achieve a prediction error of 19.6 % for encoding time and 20.9 % for encoding energy.
This paper presents LatentPatch, a new method for generating realistic images from a small dataset of only a few images. We use a lightweight model with only a few thousand parameters. Unlike traditional few-shot generation methods that finetune pre-trained large-scale generative models, our approach is computed directly on the latent distribution by sequential feature matching, and is explainable by design. Avoiding large models based on transformers, recursive networks, or self-attention, which are not suitable for small datasets, our method is inspired by non-parametric texture synthesis and style transfer models, and ensures that generated image features are sampled from the source distribution. We extend previous single-image models to work with a few images and demonstrate that our method can generate realistic images, as well as enable conditional sampling and image editing. We conduct experiments on face datasets and show that our simplistic model is effective and versatile.
To reconstruct a 3D human surface from a single image, it is important to consider human pose, shape and clothing details simultaneously. In recent years, a combination of parametric body models (such as SMPL) that capture body pose and shape prior, and neural implicit functions that learn flexible clothing details, has been used to integrate the advantages of both approaches. However, the combined representation introduces additional computation, e.g. signed distance calculation, in 3D body feature extraction, which exacerbates the redundancy of the implicit query-and-infer process and fails to preserve the underlying body shape prior. To address these issues, we propose a novel IUVD-Feedback representation, which consists of an IUVD occupancy function and a feedback query algorithm. With this representation, the time-consuming signed distance calculation is replaced by a simple linear transformation in the IUVD space, leveraging the SMPL UV maps. Additionally, the redundant query points in the query-and-infer process are reduced through a feedback mechanism. This leads to more reasonable 3D body features and more effective query points, successfully preserving the parametric body prior. Moreover, the IUVD-Feedback representation can be embedded into any existing implicit human reconstruction pipelines without modifying the trained neural networks. Experiments on THuman2.0 dataset demonstrate that the proposed IUVD-Feedback representation improves result robustness and achieves three times faster acceleration in the query-and-infer process. Furthermore, this representation has the potential to be used in generative applications by leveraging its inherited semantic information from the parametric body model.
Apparel's significant role in human appearance underscores the importance of garment digitalization for digital human creation. Recent advances in 3D content creation are pivotal for digital human creation. Nonetheless, garment generation from text guidance is still nascent. We introduce a text-driven 3D garment generation framework, DressCode, which aims to democratize design for novices and offer immense potential in fashion design, virtual try-on, and digital human creation. For our framework, we first introduce SewingGPT, a GPT-based architecture integrating cross-attention with text-conditioned embedding to generate sewing patterns with text guidance. We also tailored a pre-trained Stable Diffusion for high-quality, tile-based PBR texture generation. By leveraging a large language model, our framework generates CG-friendly garments through natural language interaction. Our method also facilitates pattern completion and texture editing, simplifying the process for designers by user-friendly interaction. With comprehensive evaluations and comparisons with other state-of-the-art methods, our method showcases the best quality and alignment with input prompts. User studies further validate our high-quality rendering results, highlighting its practical utility and potential in production settings.
Recently, text-to-image diffusion models have demonstrated impressive ability to generate high-quality images conditioned on the textual input. However, these models struggle to accurately adhere to textual instructions regarding spatial layout information. While previous research has primarily focused on aligning cross-attention maps with layout conditions, they overlook the impact of the initialization noise on the layout guidance. To achieve better layout control, we propose leveraging a spatial-aware initialization noise during the denoising process. Specifically, we find that the inverted reference image with finite inversion steps contains valuable spatial awareness regarding the object's position, resulting in similar layouts in the generated images. Based on this observation, we develop an open-vocabulary framework to customize a spatial-aware initialization noise for each layout condition. Without modifying other modules except the initialization noise, our approach can be seamlessly integrated as a plug-and-play module within other training-free layout guidance frameworks. We evaluate our approach quantitatively and qualitatively on the available Stable Diffusion model and COCO dataset. Equipped with the spatial-aware latent initialization, our method significantly improves the effectiveness of layout guidance while preserving high-quality content.
In HTTP adaptive live streaming applications, video segments are encoded at a fixed set of bitrate-resolution pairs known as bitrate ladder. Live encoders use the fastest available encoding configuration, referred to as preset, to ensure the minimum possible latency in video encoding. However, an optimized preset and optimized number of CPU threads for each encoding instance may result in (i) increased quality and (ii) efficient CPU utilization while encoding. For low latency live encoders, the encoding speed is expected to be more than or equal to the video framerate. To this light, this paper introduces a Just Noticeable Difference (JND)-Aware Low latency Encoding Scheme (JALE), which uses random forest-based models to jointly determine the optimized encoder preset and thread count for each representation, based on video complexity features, the target encoding speed, the total number of available CPU threads, and the target encoder. Experimental results show that, on average, JALE yield a quality improvement of 1.32 dB PSNR and 5.38 VMAF points with the same bitrate, compared to the fastest preset encoding of the HTTP Live Streaming (HLS) bitrate ladder using x265 HEVC open-source encoder with eight CPU threads used for each representation. These enhancements are achieved while maintaining the desired encoding speed. Furthermore, on average, JALE results in an overall storage reduction of 72.70 %, a reduction in the total number of CPU threads used by 63.83 %, and a 37.87 % reduction in the overall encoding time, considering a JND of six VMAF points.
Document AI is a growing research field that focuses on the comprehension and extraction of information from scanned and digital documents to make everyday business operations more efficient. Numerous downstream tasks and datasets have been introduced to facilitate the training of AI models capable of parsing and extracting information from various document types such as receipts and scanned forms. Despite these advancements, both existing datasets and models fail to address critical challenges that arise in industrial contexts. Existing datasets primarily comprise short documents consisting of a single page, while existing models are constrained by a limited maximum length, often set at 512 tokens. Consequently, the practical application of these methods in financial services, where documents can span multiple pages, is severely impeded. To overcome these challenges, we introduce LongFin, a multimodal document AI model capable of encoding up to 4K tokens. We also propose the LongForms dataset, a comprehensive financial dataset that encapsulates several industrial challenges in financial documents. Through an extensive evaluation, we demonstrate the effectiveness of the LongFin model on the LongForms dataset, surpassing the performance of existing public models while maintaining comparable results on existing single-page benchmarks.
As a scene graph compactly summarizes the high-level content of an image in a structured and symbolic manner, the similarity between scene graphs of two images reflects the relevance of their contents. Based on this idea, we propose a novel approach for image-to-image retrieval using scene graph similarity measured by graph neural networks. In our approach, graph neural networks are trained to predict the proxy image relevance measure, computed from human-annotated captions using a pre-trained sentence similarity model. We collect and publish the dataset for image relevance measured by human annotators to evaluate retrieval algorithms. The collected dataset shows that our method agrees well with the human perception of image similarity than other competitive baselines.
Deep neural networks (DNNs) are successful in many computer vision tasks. However, the most accurate DNNs require millions of parameters and operations, making them energy, computation and memory intensive. This impedes the deployment of large DNNs in low-power devices with limited compute resources. Recent research improves DNN models by reducing the memory requirement, energy consumption, and number of operations without significantly decreasing the accuracy. This paper surveys the progress of low-power deep learning and computer vision, specifically in regards to inference, and discusses the methods for compacting and accelerating DNN models. The techniques can be divided into four major categories: (1) parameter quantization and pruning, (2) compressed convolutional filters and matrix factorization, (3) network architecture search, and (4) knowledge distillation. We analyze the accuracy, advantages, disadvantages, and potential solutions to the problems with the techniques in each category. We also discuss new evaluation metrics as a guideline for future research.
Image segmentation is considered to be one of the critical tasks in hyperspectral remote sensing image processing. Recently, convolutional neural network (CNN) has established itself as a powerful model in segmentation and classification by demonstrating excellent performances. The use of a graphical model such as a conditional random field (CRF) contributes further in capturing contextual information and thus improving the segmentation performance. In this paper, we propose a method to segment hyperspectral images by considering both spectral and spatial information via a combined framework consisting of CNN and CRF. We use multiple spectral cubes to learn deep features using CNN, and then formulate deep CRF with CNN-based unary and pairwise potential functions to effectively extract the semantic correlations between patches consisting of three-dimensional data cubes. Effective piecewise training is applied in order to avoid the computationally expensive iterative CRF inference. Furthermore, we introduce a deep deconvolution network that improves the segmentation masks. We also introduce a new dataset and experimented our proposed method on it along with several widely adopted benchmark datasets to evaluate the effectiveness of our method. By comparing our results with those from several state-of-the-art models, we show the promising potential of our method.
小貼士
登錄享
相關主題
注冊地址： 北京市海淀區羊坊店路18號2幢3層301-191