Graphics and Human Computer Interfaces Commons^™

Toward Learning-Based Reconstruction And Part Decomposition Of Man-Made 3d Geometry: Neural Implicit Representations And Scalable Supervision, Shen Fan

Dissertations

Digital three-dimensional (3D) models are central to engineering design, analysis, and manufacturing, but learning pipelines for man-made geometry often operate on sampled carriers that do not preserve all of the structure present in exact CAD representations. This dissertation studies learning-based reconstruction and part decomposition for structured man-made 3D geometry, from general object benchmarks to CAD-derived datasets, with a focus on neural implicit representations trained from signed-distance samples, point clouds, and tessellated meshes. The goal is to make these models more accurate, more part-aware, and more consistently supervised.

First, signed distance function (SDF) reconstruction with implicit neural representations is improved through …

Spaceforge: Spatial Reconstruction For Signal Simulations, Compton Ross

Honors Theses

Signal simulation environments require accurate three dimensional representations of physical spaces, yet current methods for generating these representations, including Light Detection and Ranging (LiDAR) scanning, manual 3D modeling, and commercial photogrammetry, are both costly and time intensive. SpaceForge addresses this gap with a prompt guided pipeline that takes an ordinary indoor photograph and a configurable set of simulation relevant object categories as input and produces a voxelized 3D scene compatible with downstream signal simulation workflows. The pipeline proceeds through five major stages: open set object detection and segmentation, object level preprocessing, single image 3D mesh reconstruction, heuristic pose estimation and …

Instruct2see: Learning To Remove Any Obstructions Across Distributions, Junhang Li, Yu Guo, Chuhua Xian, Shengfeng He

Research Collection School Of Computing and Information Systems

Images are often obstructed by various obstacles due to capture limitations, hindering the observation of objects of interest. Most existing methods address occlusions from specific elements like fences or raindrops, but are constrained by the wide range of real-world obstructions, making comprehensive data collection impractical. To overcome these challenges, we propose Instruct2See, a novel zero-shot framework capable of handling both seen and unseen obstacles. The core idea of our approach is to unify obstruction removal by treating it as a soft-hard mask restoration problem, where any obstruction can be represented using multi-modal prompts, such as visual semantics and textual instructions, …

Human Perception Of Ai Capabilities At Classifying Perturbed Roadway Signs, Katherine R. Garcia, Jing Chen, Yanru Xiao, Scott Mishler, Cong Wang, Bin Hu

Computer Science Faculty Publications

Artificial Intelligence (AI) is crucial to numerous functions required for driving automation systems, including the computer vision techniques used to detect the roadway environment and make real-time decisions. However, the images used as inputs to the AI system may be maliciously perturbed, or manipulated, causing the AI system to make an incorrect classification. In this study, we examined humans’ perception of the AI’s computer vision capability of classifying various road sign images, including the original images, images with two different types of malicious attacks, and images that are scrambled randomly at the pixel level. Our results showed that participants rated …

Fast And High-Resolution View Synthesis From A Single Input Panorama, Nam Nguyen, Angela V. Chen, Theresa Zhu, Seth Johnson, Pranav Dumpa, Benjamin Geil

College of Engineering Summer Undergraduate Research Program

We introduce a novel method to convert a single input panorama into a 3D colored mesh representation of the scene. Unlike recent methods based on neural rendering, which are limited to low-resolution inputs and offline rendering, our approach supports 4k resolution inputs and real time rendering in a virtual reality headset. We first estimate a depth map and produce an initial layered depth image (LDI) representation. We fill unseen regions behind objects by iteratively cutting and inpainting the LDI. We then convert the LDI into an optimized, texture mapped mesh to achieve a compact representation

Improving Out-Of-Distribution Detection With Disentangled Foreground And Background Features, Choubo Ding, Guansong Pang

Research Collection School Of Computing and Information Systems

Detecting out-of-distribution (OOD) inputs is a principal task for ensuring the safety of deploying deep-neural-network classifiers in open-set scenarios. OOD samples can be drawn from arbitrary distributions and exhibit deviations from in-distribution (ID) data in various dimensions, such as foreground features (e.g., objects in CIFAR100 images vs. those in CIFAR10 images) and background features (e.g., textural images vs. objects in CIFAR10). Existing methods can confound foreground and background features in training, failing to utilize the background features for OOD detection. This paper considers the importance of feature disentanglement in out-of-distribution detection and proposes the simultaneous exploitation of both foreground and …

Beyond Textual Constraints : Learning Novel Diffusion Conditions With Fewer Examples, Yuyang Yu, Bangzhen Liu, Chenxi Zheng, Xuemiao Xu, Huaidong Zhang, Shengfeng He

Research Collection School Of Computing and Information Systems

In this paper, we delve into a novel aspect of learning novel diffusion conditions with datasets an order of magnitude smaller. The rationale behind our approach is the elimination of textual constraints during the few-shot learning process. To that end, we implement two optimization strategies. The first, prompt-free conditional learning, utilizes a prompt-free encoder derived from a pre-trained Stable Diffusion model. This strategy is designed to adapt new conditions to the diffusion process by minimizing the textual-visual cor-relation, thereby ensuring a more precise alignment between the generated content and the specified conditions. The second strategy entails condition-specific negative rectification, which …

Learning With Unreliability : Fast Few-Shot Voxel Radiance Fields With Relative Geometric Consistency, Yingjie Xu, Bangzhen Liu, Hao Tang, Bailin Deng, Shengfeng He

Research Collection School Of Computing and Information Systems

We propose a voxel-based optimization framework, Re VoRF, for few-shot radiance fields that strategically ad-dress the unreliability in pseudo novel view synthesis. Our method pivots on the insight that relative depth relationships within neighboring regions are more reliable than the ab-solute color values in disoccluded areas. Consequently, we devise a bilateral geometric consistency loss that carefully navigates the trade-off between color fidelity and geometric accuracy in the context of depth consistency for uncertain regions. Moreover, we present a reliability-guided learning strategy to discern and utilize the variable quality across syn-thesized views, complemented by a reliability-aware voxel smoothing algorithm that smoothens …

Rethinking Multi-View Representation Learning Via Distilled Disentangling, Guanzhou Ke, Bo Wang, Xiaoli Wang, Shengfeng He

Research Collection School Of Computing and Information Systems

Multi-view representation learning aims to derive robust representations that are both view-consistent and view-specific from diverse data sources. This paper presents an in-depth analysis of existing approaches in this domain, highlighting a commonly overlooked aspect: the redundancy between view-consistent and view-specific representations. To this end, we propose an innovative framework for multi-view representation learning, which incorporates a technique we term 'distilled disentangling'. Our method introduces the concept of masked cross-view prediction, enabling the extraction of compact, high-quality view-consistent representations from various sources without incurring extra computational overhead. Additionally, we develop a distilled disentangling module that efficiently filters out consistency-related information …

Embodied Visions: Interactive Installations That Reimagine Bodily Presence In Digital Imaging Apparatuses As Shadows, Yunzi Shi

Dartmouth College Master’s Theses

Contextualized within a history of technological development, the evolution of imaging devices and technologies is accompanied by the abstraction of spatial relationships between the body of the observer, the apparatus, and physical reality, which leads to disembodying experiences for the observing subject. Compared with devices and interactive experiences, critical reflection on the epistemological impact of digital imaging devices has less priority in computational imaging and human-computer interaction research. Taking an artistic approach, this thesis describes Embodied Visions, an exhibition featuring three interactive installations exploring the technical infrastructure for imaging and reflecting on the (dis)embodied experiences in the digital age. …

3-D Reconstruction For Underwater Robots With A Monocular Camera And Lights, Monika Roznere

Dartmouth College Ph.D Dissertations

Before a robot can act, it must perceive its environment. Though, this is not a simple task when considering the challenges in underwater domains -- poor visibility conditions, limited sensor configurations, and lack of readily accessible localization. Underwater robots have, nevertheless, improved dramatically with more extensive sensor and navigation equipment. Robot and sensor use have enabled us to explore all reaches of our oceans. On the other hand, these same robots are not easily accessible or transferable to many practical tasks, including fishery management, infrastructure maintenance, disaster response, site conservation, and ecological surveys. There is a growing need for robots …

Deep Learning Image Analysis To Isolate And Characterize Different Stages Of S-Phase In Human Cells, Kevin A. Boyd, Rudranil Mitra, John Santerre, Christopher L. Sansam

SMU Data Science Review

Abstract. This research used deep learning for image analysis by isolating and characterizing distinct DNA replication patterns in human cells. By leveraging high-resolution microscopy images of multiple cells stained with 5-Ethynyl-2′-deoxyuridine (EdU), a replication marker, this analysis utilized Convolutional Neural Networks (CNNs) to perform image segmentation and to provide robust and reliable classification results. First multiple cells in a field of focus were identified using a pretrained CNN called Cellpose. After identifying the location of each cell in the image a python script was created to crop out each cell into individual .tif files. After careful annotation, a CNN was …

Enhanced Privacy-Enabled Face Recognition Using Κ-Identity Optimization, Ryan Karl

Department of Electrical and Computer Engineering: Dissertations, Theses, and Student Research

Facial recognition is becoming more and more prevalent in the daily lives of the common person. Law enforcement utilizes facial recognition to find and track suspects. The newest smartphones have the ability to unlock using the user's face. Some door locks utilize facial recognition to allow correct users to enter restricted spaces. The list of applications that use facial recognition will only increase as hardware becomes more cost-effective and more computationally powerful. As this technology becomes more prevalent in our lives, it is important to understand and protect the data provided to these companies. Any data transmitted should be encrypted …

Stprivacy: Spatio-Temporal Privacy-Preserving Action Recognition, Ming Li, Xiangyu Xu, Hehe Fan, Pan Zhou, Jun Liu, Jia-Wei Liu, Jiahe Li, Jussi Keppo, Mike Zheng Shou, Shuicheng Yan

Research Collection School Of Computing and Information Systems

Existing methods of privacy-preserving action recognition (PPAR) mainly focus on frame-level (spatial) privacy removal through 2D CNNs. Unfortunately, they have two major drawbacks. First, they may compromise temporal dynamics in input videos, which are critical for accurate action recognition. Second, they are vulnerable to practical attacking scenarios where attackers probe for privacy from an entire video rather than individual frames. To address these issues, we propose a novel framework STPrivacy to perform video-level PPAR. For the first time, we introduce vision Transformers into PPAR by treating a video as a tubelet sequence, and accordingly design two complementary mechanisms, i.e., sparsification …

Fine-Grained Domain Adaptive Crowd Counting Via Point-Derived Segmentation, Yongtuo Liu, Dan Xu, Sucheng Ren, Hanjie Wu, Hongmin Cai, Shengfeng He

Research Collection School Of Computing and Information Systems

Due to domain shift, a large performance drop is usually observed when a trained crowd counting model is deployed in the wild. While existing domain-adaptive crowd counting methods achieve promising results, they typically regard each crowd image as a whole and reduce domain discrepancies in a holistic manner, thus limiting further improvement of domain adaptation performance. To this end, we propose to untangle domain-invariant crowd and domain-specific background from crowd images and design a fine-grained domain adaption method for crowd counting. Specifically, to disentangle crowd from background, we propose to learn crowd segmentation from point-level crowd counting annotations in a …

Curricular Contrastive Regularization For Physics-Aware Single Image Dehazing, Yu Zheng, Jiahui Zhan, Shengfeng He, Yong Du

Research Collection School Of Computing and Information Systems

Considering the ill-posed nature, contrastive regularization has been developed for single image dehazing, introducing the information from negative images as a lower bound. However, the contrastive samples are non-consensual, as the negatives are usually represented distantly from the clear (i.e., positive) image, leaving the solution space still under-constricted. Moreover, the interpretability of deep dehazing models is underexplored towards the physics of the hazing process. In this paper, we propose a novel curricular contrastive regularization targeted at a consensual contrastive space as opposed to a non-consensual one. Our negatives, which provide better lower-bound constraints, can be assembled from 1) the hazy …

High-Resolution Face Swapping Via Latent Semantics Disentanglement, Yangyang Xu, Bailin Deng, Junle Wang, Yanqing Jing, Jia Pan, Shengfeng He

Research Collection School Of Computing and Information Systems

We present a novel high-resolution face swapping method using the inherent prior knowledge of a pre-trained GAN model. Although previous research can leverage generative priors to produce high-resolution results, their quality can suffer from the entangled semantics of the latent space. We explicitly disentangle the latent semantics by utilizing the progressive nature of the generator, deriving structure at-tributes from the shallow layers and appearance attributes from the deeper ones. Identity and pose information within the structure attributes are further separated by introducing a landmark-driven structure transfer latent direction. The disentangled latent code produces rich generative features that incorporate feature blending …

Metaformer Is Actually What You Need For Vision, Weihao Yu, Mi Luo, Pan Zhou, Chenyang Si, Yichen Zhou, Xinchao Wang, Jiashi Feng, Shuicheng Yan

Research Collection School Of Computing and Information Systems

Transformers have shown great potential in computer vision tasks. A common belief is their attention-based token mixer module contributes most to their competence. However, recent works show the attention-based module in transformers can be replaced by spatial MLPs and the resulted models still perform quite well. Based on this observation, we hypothesize that the general architecture of the transformers, instead of the specific token mixer module, is more essential to the model's performance. To verify this, we deliberately replace the attention module in transformers with an embarrassingly simple spatial pooling operator to conduct only basic token mixing. Surprisingly, we observe …

Gauging The State-Of-The-Art For Foresight Weight Pruning On Neural Networks, Noah James

Computer Science and Computer Engineering Undergraduate Honors Theses

The state-of-the-art for pruning neural networks is ambiguous due to poor experimental practices in the field. Newly developed approaches rarely compare to each other, and when they do, their comparisons are lackluster or contain errors. In the interest of stabilizing the field of pruning, this paper initiates a dive into reproducing prominent pruning algorithms across several architectures and datasets. As a first step towards this goal, this paper shows results for foresight weight pruning across 6 baseline pruning strategies, 5 modern pruning strategies, random pruning, and one legacy method (Optimal Brain Damage). All strategies are evaluated on 3 different architectures …

Real Time Evaluation Of Boom And Drogue Occlusion With Aar, Xiaoyang Wu

Theses and Dissertations

In recent years, Unmanned Aerial Vehicles (UAV) have seen a rise in popularity. Various navigational algorithms have been developed as a solution to estimate a UAV’s pose relative to the refueler aircraft. The result can be used to safely automate aerial refueling (AAR) to improve UAVs’ time-on-station and ensure the success of military operations. This research aims to reach real-time performance using a GPU accelerated approach. It also conducts various experiments to quantify the effects of refueling boom/drogue occlusion and image exposure on the pose estimation pipeline in a lab setting.

Projecting Your View Attentively: Monocular Road Scene Layout Estimation Via Cross-View Transformation, Weixiang Yang, Qi Li, Wenxi Liu, Yuanlong Yu, Yuexin Ma, Shengfeng He, Jia Pan

Research Collection School Of Computing and Information Systems

HD map reconstruction is crucial for autonomous driving. LiDAR-based methods are limited due to the deployed expensive sensors and time-consuming computation. Camera-based methods usually need to separately perform road segmentation and view transformation, which often causes distortion and the absence of content. To push the limits of the technology, we present a novel framework that enables reconstructing a local map formed by road layout and vehicle occupancy in the bird's-eye view given a front-view monocular image only. In particular, we propose a cross-view transformation module, which takes the constraint of cycle consistency between views into account and makes full use …

Reciprocal Transformations For Unsupervised Video Object Segmentation, Sucheng Ren, Wenxi Liu, Yongtuo Liu, Haoxin Chen, Guoqiang Han, Shengfeng He

Research Collection School Of Computing and Information Systems

Unsupervised video object segmentation (UVOS) aims at segmenting the primary objects in videos without any human intervention. Due to the lack of prior knowledge about the primary objects, identifying them from videos is the major challenge of UVOS. Previous methods often regard the moving objects as primary ones and rely on optical flow to capture the motion cues in videos, but the flow information alone is insufficient to distinguish the primary objects from the background objects that move together. This is because, when the noisy motion features are combined with the appearance features, the localization of the primary objects is …

Fakepolisher: Making Deepfakes More Detection-Evasive By Shallow Reconstruction, Yihao Huang, Felix Juefei-Xu, Run Wang, Qing Guo, Lei Ma, Xiaofei Xie, Jianwen Li, Weikai Miao, Yang Liu, Geguang Pu

Research Collection School Of Computing and Information Systems

At this moment, GAN-based image generation methods are still imperfect, whose upsampling design has limitations in leaving some certain artifact patterns in the synthesized image. Such artifact patterns can be easily exploited (by recent methods) for difference detection of real and GAN-synthesized images. However, the existing detection methods put much emphasis on the artifact patterns, which can become futile if such artifact patterns were reduced.Towards reducing the artifacts in the synthesized images, in this paper, we devise a simple yet powerful approach termed FakePolisher that performs shallow reconstruction of fake images through a learned linear dictionary, intending to effectively and …

Computer Vision-Based Traffic Sign Detection And Extraction: A Hybrid Approach Using Gis And Machine Learning, Zihao Wu

College of Graduate Studies: Theses & Dissertations

Traffic sign detection and positioning have drawn considerable attention because of the recent development of autonomous driving and intelligent transportation systems. In order to detect and pinpoint traffic signs accurately, this research proposes two methods. In the first method, geo-tagged Google Street View images and road networks were utilized to locate traffic signs. In the second method, both traffic signs categories and locations were identified and extracted from the location-based GoPro video. TensorFlow is the machine learning framework used to implement these two methods. To that end, 363 stop signs were detected and mapped accurately using the first method (Google …

Formresnet: Formatted Residual Learning For Image Restoration, Jianbo Jiao, Wei-Chih Tu, Shengfeng He

Research Collection School Of Computing and Information Systems

In this paper, we propose a deep CNN to tackle the image restoration problem by learning the structured residual. Previous deep learning based methods directly learn the mapping from corrupted images to clean images, and may suffer from the gradient exploding/vanishing problems of deep neural networks. We propose to address the image restoration problem by learning the structured details and recovering the latent clean image together, from the shared information between the corrupted image and the latent image. In addition, instead of learning the pure difference (corruption), we propose to add a 'residual formatting layer' to format the residual to …

Real-Time, Non-Contact Heart Rate Monitor, Daniel Blike

Computer Engineering

No abstract provided.

Exemplar-Driven Top-Down Saliency Detection Via Deep Association, Shengfeng He, Rynson W. H. Lau, Qingxiong Yang

Research Collection School Of Computing and Information Systems

Top-down saliency detection is a knowledge-driven search task. While some previous methods aim to learn this "knowledge" from category-specific data, others transfer existing annotations in a large dataset through appearance matching. In contrast, we propose in this paper a locateby-exemplar strategy. This approach is challenging, as we only use a few exemplars (up to 4) and the appearances among the query object and the exemplars can be very different. To address it, we design a two-stage deep model to learn the intra-class association between the exemplars and query objects. The first stage is for learning object-to-object association, and the second …

Un Indicator De Incluziune Cu Aplicaţii În Computer Vision, Florentin Smarandache, Ovidiu Ilie Sandru

Branch Mathematics and Statistics Faculty and Staff Publications

În aceasta lucrare vom prezenta un procedeu de algoritmizare a operatiilor necesare deplasarii automate a unui obiect predefinit dintr-o imagine video data intr-o regiune tinta a acelei imagini, menit a facilita realizarea de aplicatii software specializate in rezolvarea acestui gen de probleme.

Oriented Object Proposals, Shengfeng He, Rynson W. H. Lau

Research Collection School Of Computing and Information Systems

In this paper, we propose a new approach to generate oriented object proposals (OOPs) to reduce the detection error caused by various orientations of the object. To this end, we propose to efficiently locate object regions according to pixelwise object probability, rather than measuring the objectness from a set of sampled windows. We formulate the proposal generation problem as a generative probabilistic model such that object proposals of different shapes (i.e., sizes and orientations) can be produced by locating the local maximum likelihoods. The new approach has three main advantages. First, it helps the object detector handle objects of different …

Designing A Bayer Filter With Smooth Hue Transition Interpolation Using The Xilinx System Generator, Zhiqiang Li, Peter Revesz

School of Computing: Conference and Workshop Papers

This paper describes the design of a Bayer filter with smooth hue transition using the System Generator for DSP. We describe and compare experimentally two different designs, one based on a MATLAB implementation and the other based on a modification of the Bayer filter using bilinear interpolation.