Search

We present EucliDreamer, a simple and effective method to generate textures for 3D models given text prompts and meshes. The texture is parametrized as an implicit function on the 3D surface, which is optimized with the Score Distillation Sampling (SDS) process and differentiable rendering. To generate high-quality textures, we leverage a depth-conditioned Stable Diffusion model guided by the depth image rendered from the mesh. We test our approach on 3D models in Objaverse and conducted a user study, which shows its superior quality compared to existing texturing methods like Text2Tex. In addition, our method converges 2 times faster than DreamFusion. Through text prompting, textures of diverse art styles can be produced. We hope Euclidreamer proides a viable solution to automate a labor-intensive stage in 3D content creation., Comment: Short version of arXiv:2311.15573

We propose SAM-Road, an adaptation of the Segment Anything Model (SAM) for extracting large-scale, vectorized road network graphs from satellite imagery. To predict graph geometry, we formulate it as a dense semantic segmentation task, leveraging the inherent strengths of SAM. The image encoder of SAM is fine-tuned to produce probability masks for roads and intersections, from which the graph vertices are extracted via simple non-maximum suppression. To predict graph topology, we designed a lightweight transformer-based graph neural network, which leverages the SAM image embeddings to estimate the edge existence probabilities between vertices. Our approach directly predicts the graph vertices and edges for large regions without expensive and complex post-processing heuristics, and is capable of building complete road network graphs spanning multiple square kilometers in a matter of seconds. With its simple, straightforward, and minimalist design, SAM-Road achieves comparable accuracy with the state-of-the-art method RNGDet++, while being 40 times faster on the City-scale dataset. We thus demonstrate the power of a foundational vision model when applied to a graph learning task. The code is available at https://github.com/htcr/sam_road., Comment: Accepted by IEEE/CVF Computer Vision and Pattern Recognition Conference (CVPR) 2024, 2nd Workshop on Scene Graphs and Graph Representation Learning

This paper presents a novel method to generate textures for 3D models given text prompts and 3D meshes. Additional depth information is taken into account to perform the Score Distillation Sampling (SDS) process with depth conditional Stable Diffusion. We ran our model over the open-source dataset Objaverse and conducted a user study to compare the results with those of various 3D texturing methods. We have shown that our model can generate more satisfactory results and produce various art styles for the same object. In addition, we achieved faster time when generating textures of comparable quality. We also conduct thorough ablation studies of how different factors may affect generation quality, including sampling steps, guidance scale, negative prompts, data augmentation, elevation range, and alternatives to SDS.

Most face identification approaches employ a Siamese neural network to compare two images at the image embedding level. Yet, this technique can be subject to occlusion (e.g. faces with masks or sunglasses) and out-of-distribution data. DeepFace-EMD (Phan et al. 2022) reaches state-of-the-art accuracy on out-of-distribution data by first comparing two images at the image level, and then at the patch level. Yet, its later patch-wise re-ranking stage admits a large $O(n^3 \log n)$ time complexity (for $n$ patches in an image) due to the optimal transport optimization. In this paper, we propose a novel, 2-image Vision Transformers (ViTs) that compares two images at the patch level using cross-attention. After training on 2M pairs of images on CASIA Webface (Yi et al. 2014), our model performs at a comparable accuracy as DeepFace-EMD on out-of-distribution data, yet at an inference speed more than twice as fast as DeepFace-EMD (Phan et al. 2022). In addition, via a human study, our model shows promising explainability through the visualization of cross-attention. We believe our work can inspire more explorations in using ViTs for face identification., Comment: 20 pages, 15 Figures

Since June 2019, the Articulation of Credit Transfer project (ACT) has focused on streamlining the information, advising, and administrative processes concerning how credits from one City University of New York (CUNY) institution transfer to another CUNY institution. This work addresses a critical challenge at CUNY, and indeed, across American higher education. When students transfer from one college to another, they frequently are unable to count their previously earned credits toward degree requirements at their new institution, jeopardizing their ability to earn that degree. This report describes one of the key innovations of the ACT project: instituting a daily archive of Degree Works data for students transferring between CUNY institutions participating in the project. Creating this archive has allowed the project team and other analysts to determine changes in how these students' transferred courses counted toward their degree programs at their new institution. This report contextualizes and explains the degree audit archiving effort and how the ACT team has used the resulting data.

In the nearly two years since the election of U.S. President Donald Trump, multiple studies have documented declines in international student enrollments at U.S. institutions. Entities that track developments in higher education have reported increased enrollments in other countries, especially those where costs are lower, quality and access are on the rise, and social and political environments are more welcoming. To better understand how these shifts affect enrollments, applications, and international recruiting at U.S. institutions, the World Education Services research team surveyed more than 270 higher education professionals in early 2018. The goal was to gain evidence-based insights into changes in enrollment patterns and to develop a set of practical recommendations. Key findings include: (1) Enrollments and applications are down on the majority of campuses; (2) Most respondents felt that the political environment is a cause of their international recruitment challenges; and (3) More than a quarter of respondents reported a year-over-year increase in international applications between 2016-17 and 2017-18. Recommendations include: (1) Diversify recruitment targets; (2) Emphasize a welcoming environment; (3) Provide support and resources; (4) Engage in virtual and social media outreach; (5) Engage the alumni network for recruitment; (6) Address students' financial resources; and (7) Develop and maintain partnerships that can either provide alternative routes to recruitment or facilitate entry.

Very deep convolutional neural networks (CNNs) have been firmly established as the primary methods for many computer vision tasks. However, most state-of-the-art CNNs are large, which results in high inference latency. Recently, depth-wise separable convolution has been proposed for image recognition tasks on computationally limited platforms such as robotics and self-driving cars. Though it is much faster than its counterpart, regular convolution, accuracy is sacrificed. In this paper, we propose a novel decomposition approach based on SVD, namely depth-wise decomposition, for expanding regular convolutions into depthwise separable convolutions while maintaining high accuracy. We show our approach can be further generalized to the multi-channel and multi-layer cases, based on Generalized Singular Value Decomposition (GSVD) [59]. We conduct thorough experiments with the latest ShuffleNet V2 model [47] on both random synthesized dataset and a large-scale image recognition dataset: ImageNet [10]. Our approach outperforms channel decomposition [73] on all datasets. More importantly, our approach improves the Top-1 accuracy of ShuffleNet V2 by ~2%.

Each year, the country's most selective four-year institutions invest significant resources to recruit talented high school students from across the country. These students, mostly affluent and white, contemplate admission offers and consider moves to new locales to pursue their postsecondary plans. Yet, many of these selective institutions are overlooking a talented and diverse pool of students in their own backyard: transfer students from local community colleges. In this issue brief, the authors set out to answer the following question: What is the supply of qualified community college students in the local area of our nation's high-graduation-rate colleges and universities? The authors define "high-graduation-rate" as those 334 public and private, not-for-profit colleges and universities with six-year graduation rates consistently above 70 percent. The authors also provide examples of two high-graduation-rate colleges, both privates, that have developed partnerships with their local community colleges to expand transfer enrollment to their campuses. These examples demonstrate the different forms these partnerships can take depending on the needs of both the two- and four-year colleges and their students. They serve as a model for other private institutions that aim to expand transfer pipelines, diversity student enrollment, and improve student outcomes.