Matching Items (61)
Filtering by
- Genre: Masters Thesis
- Creators: Yang, Yezhou
- Resource Type: Text
Description
Simultaneous localization and mapping (SLAM) has traditionally relied on low-level geometric or optical features. However, these features-based SLAM methods often struggle with feature-less or repetitive scenes. Additionally, low-level features may not provide sufficient information for robot navigation and manipulation, leaving robots without a complete understanding of the 3D spatial world. Advanced information is necessary to address these limitations. Fortunately, recent developments in learning-based 3D reconstruction allow robots to not only detect semantic meanings, but also recognize the 3D structure of objects from a few images. By combining this 3D structural information, SLAM can be improved from a low-level approach to a structure-aware approach. This work propose a novel approach for multi-view 3D reconstruction using recurrent transformer. This approach allows robots to accumulate information from multiple views and encode them into a compact latent space. The resulting latent representations are then decoded to produce 3D structural landmarks, which can be used to improve robot localization and mapping.
ContributorsHuang, Chi-Yao (Author) / Yang, Yezhou (Thesis advisor) / Turaga, Pavan (Committee member) / Jayasuriya, Suren (Committee member) / Arizona State University (Publisher)
Created2023
Description
In recent years, there has been significant progress in deep learning and computer vision, with many models proposed that have achieved state-of-art results on various image recognition tasks. However, to explore the full potential of the advances in this field, there is an urgent need to push the processing of deep networks from the cloud to edge devices. Unfortunately, many deep learning models cannot be efficiently implemented on edge devices as these devices are severely resource-constrained. In this thesis, I present QU-Net, a lightweight binary segmentation model based on the U-Net architecture. Traditionally, neural networks consider the entire image to be significant. However, in real-world scenarios, many regions in an image do not contain any objects of significance. These regions can be removed from the original input allowing a network to focus on the relevant regions and thus reduce computational costs. QU-Net proposes the salient regions (binary mask) that the deeper models can use as the input. Experiments show that QU-Net helped achieve a computational reduction of 25% on the Microsoft Common Objects in Context (MS COCO) dataset and 57% on the Cityscapes dataset. Moreover, QU-Net is a generalizable model that outperforms other similar works, such as Dynamic Convolutions.
ContributorsSanthosh Kumar Varma, Rahul (Author) / Yang, Yezhou (Thesis advisor) / Fan, Deliang (Committee member) / Yang, Yingzhen (Committee member) / Arizona State University (Publisher)
Created2021
Description
Generative models in various domain such as images, speeches, and videos are beingdeveloped actively over the last decades and recent deep generative models are now
capable of synthesizing multimedia contents are difficult to be distinguishable from
authentic contents. Such capabilities cause concerns such as malicious impersonation,
Intellectual property theft(IP theft) and copyright infringement.
One method to solve these threats is to embedded attributable watermarking in
synthesized contents so that user can identify the user-end models where the contents
are generated from. This paper investigates a solution for model attribution, i.e., the
classification of synthetic contents by their source models via watermarks embedded
in the contents. Existing studies showed the feasibility of model attribution in the
image domain and tradeoff between attribution accuracy and generation quality under
the various adversarial attacks but not in speech domain.
This work discuss the feasibility of model attribution in different domain and
algorithmic improvements for generating user-end speech models that empirically
achieve high accuracy of attribution while maintaining high generation quality. Lastly,
several experiments are conducted show the tradeoff between attributability and
generation quality under a variety of attacks on generated speech signals attempting
to remove the watermarks.
ContributorsCho, Yongbaek (Author) / Yang, Yezhou (Thesis advisor) / Ren, Yi (Committee member) / Trieu, Ni (Committee member) / Arizona State University (Publisher)
Created2021
Description
Multimodal reasoning is one of the most interesting research fields because of the ability to interact with systems and the explainability of the models' behavior. Traditional multimodal research problems do not focus on complex commonsense reasoning (such as physical interactions). Although real-world objects have physical properties associated with them, many of these properties (such as mass and coefficient of friction) are not captured directly by the imaging pipeline. Videos often capture objects, their motion, and the interactions between different objects. However, these properties can be estimated by utilizing cues from relative object motion and the dynamics introduced by collisions. This thesis introduces a new video question-answering task for reasoning about the implicit physical properties of objects in a scene, from videos. For this task, I introduce a dataset -- CRIPP-VQA (Counterfactual Reasoning about Implicit Physical Properties - Video Question Answering), which contains videos of objects in motion, annotated with hypothetical/counterfactual questions about the effect of actions (such as removing, adding, or replacing objects), questions about planning (choosing actions to perform to reach a particular goal), as well as descriptive questions about the visible properties of objects. Further, I benchmark the performance of existing video question-answering models on two test settings of CRIPP-VQA: i.i.d. and an out-of-distribution setting which contains objects with values of mass, coefficient of friction, and initial velocities that are not seen in the training distribution. Experiments reveal a surprising and significant performance gap in terms of answering questions about implicit properties (the focus of this thesis) and explicit properties (the focus of prior work) of objects.
ContributorsPatel, Maitreya Jitendra (Author) / Yang, Yezhou (Thesis advisor) / Baral, Chitta (Committee member) / Lee, Kookjin (Committee member) / Arizona State University (Publisher)
Created2022
Description
For a system of autonomous vehicles functioning together in a traffic scene, 3Dunderstanding of participants in the field of view or surrounding is very essential for
assessing the safety operation of the involved. This problem can be decomposed into online
pose and shape estimation, which has been a core research area of computer vision for over
a decade now. This work is an add-on to support and improve the joint estimate of the pose
and shape of vehicles from monocular cameras. The objective of jointly estimating the
vehicle pose and shape online is enabled by what is called an offline reconstruction
pipeline. In the offline reconstruction step, an approach to obtain the vehicle 3D shape with
keypoints labeled is formulated.
This work proposes a multi-view reconstruction pipeline using images and masks
which can create an approximate shape of vehicles and can be used as a shape prior. Then
a 3D model-fitting optimization approach to refine the shape prior using high quality
computer-aided design (CAD) models of vehicles is developed. A dataset of such 3D
vehicles with 20 keypoints annotated is prepared and call it the AvaCAR dataset. The
AvaCAR dataset can be used to estimate the vehicle shape and pose, without having the
need to collect significant amounts of data needed for adequate training of a neural
network. The online reconstruction can use this synthesis dataset to generate novel
viewpoints and simultaneously train a neural network for pose and shape estimation. Most
methods in the current literature using deep neural networks, that are trained to estimate
pose of the object from a single image, are inherently biased to the viewpoint of the images
used. This approach aims at addressing these existing limitations in the current method by
delivering the online estimation a shape prior which can generate novel views to account
for the bias due to viewpoint. The dataset is provided with ground truth extrinsic parameters
and the compact vector based shape representations which along with the multi-view
dataset can be used to efficiently trained neural networks for vehicle pose and shape
estimation. The vehicles in this library are evaluated with some standard metrics to assure
they are capable of aiding online estimation and model based tracking.
ContributorsDUTTA, PRABAL BIJOY (Author) / Yang, Yezhou (Thesis advisor) / Berman, Spring (Committee member) / Lu, Duo (Committee member) / Arizona State University (Publisher)
Created2022
Description
Recent advances in autonomous vehicle (AV) technologies have ensured that autonomous driving will soon be present in real-world traffic. Despite the potential of AVs, many studies have shown that traffic accidents in hybrid traffic environments (where both AVs and human-driven vehicles (HVs) are present) are inevitable because of the unpredictability of human-driven vehicles. Given that eliminating accidents is impossible, an achievable goal of designing AVs is to design them in a way so that they will not be blamed for any accident in which they are involved in. This work proposes BlaFT – a Blame-Free motion planning algorithm in hybrid Traffic. BlaFT is designed to be compatible with HVs and other AVs, and will not be blamed for accidents in a structured road environment. Also, it proves that no accidents will happen if all AVs are using the BlaFT motion planner and that when in hybrid traffic, the AV using BlaFT will be blame-free even if it is involved in a collision. The work instantiated scores of BlaFT and HV vehicles in an urban road scape loop in the 'Simulation of Urban MObility', ran the simulation for several hours, and observe that as the percentage of BlaFT vehicles increases, the traffic becomes safer. Adding BlaFT vehicles to HVs also increases the efficiency of traffic as a whole by up to 34%.
ContributorsPark, Sanggu (Author) / Shrivastava, Aviral (Thesis advisor) / Wang, Ruoyu (Committee member) / Yang, Yezhou (Committee member) / Arizona State University (Publisher)
Created2022
Description
Currently, one of the biggest limiting factors for long-term deployment of autonomous systems is the power constraints of a platform. In particular, for aerial robots such as unmanned aerial vehicles (UAVs), the energy resource is the main driver of mission planning and operation definitions, as everything revolved around flight time. The focus of this work is to develop a new method of energy storage and charging for autonomous UAV systems, for use during long-term deployments in a constrained environment. We developed a charging solution that allows pre-equipped UAV system to land on top of designated charging pads and rapidly replenish their battery reserves, using a contact charging point. This system is designed to work with all types of rechargeable batteries, focusing on Lithium Polymer (LiPo) packs, that incorporate a battery management system for increased reliability. The project also explores optimization methods for fleets of UAV systems, to increase charging efficiency and extend battery lifespans. Each component of this project was first designed and tested in computer simulation. Following positive feedback and results, prototypes for each part of this system were developed and rigorously tested. Results show that the contact charging method is able to charge LiPo batteries at a 1-C rate, which is the industry standard rate, maintaining the same safety and efficiency standards as modern day direct connection chargers. Control software for these base stations was also created, to be integrated with a fleet management system, and optimizes UAV charge levels and distribution to extend LiPo battery lifetimes while still meeting expected mission demand. Each component of this project (hardware/software) was designed for manufacturing and implementation using industry standard tools, making it ideal for large-scale implementations. This system has been successfully tested with a fleet of UAV systems at Arizona State University, and is currently being integrated into an Arizona smart city environment for deployment.
ContributorsMian, Sami (Author) / Panchanathan, Sethuraman (Thesis advisor) / Berman, Spring (Committee member) / Yang, Yezhou (Committee member) / McDaniel, Troy (Committee member) / Arizona State University (Publisher)
Created2018
Description
I present my work on a scalable and programmable I/O controller for region-based computing, which will be used in a rhythmic pixel-based camera pipeline. I provide a breakdown of the development and design of the I/O controller and how it fits in to rhythmic pixel regions, along with a studyon memory traffic of rhythmic pixel regions and how this translates to energy efficiency. This rhythmic pixel region-based camera pipeline has been jointly developed through Dr. Robert LiKamWa’s research lab. High spatiotemporal resolutions allow high precision for vision applications, such as for detecting features for augmented reality or face detection. High spatiotemporal resolution also comes with high memory throughput, leading to higher energy usage. This creates a tradeoff between high precision and energy efficiency, which becomes more important in mobile systems. In addition, not all pixels in a frame are necessary for the vision application, such as pixels that make up the background. Rhythmic pixel regions aim to reduce the tradeoff by creating a pipeline that allows an application developer to specify regions to capture at a non-uniform spatiotemporal resolution. This is accomplished by encoding the incoming image, and only sending the pixels within these specified regions. Later these encoded representations will be decoded to a standard frame representation usable by traditional vision applications. My contribution to this effort has been the design, testing and evaluation of the I/O controller.
ContributorsNguyen, Van (Author) / LiKamWa, Robert (Thesis advisor) / Jayasuriya, Suren (Committee member) / Yang, Yezhou (Committee member) / Arizona State University (Publisher)
Created2020
Description
Insufficient training data poses significant challenges to training a deep convolutional neural network (CNN) to solve a target task. One common solution to this problem is to use transfer learning with pre-trained networks to apply knowledge learned from one domain with sufficient data to a new domain with limited data and avoid training a deep network from scratch. However, for such methods to work in a transfer learning setting, learned features from the source domain need to be generalizable to the target domain, which is not guaranteed since the feature space and distributions of the source and target data may be different. This thesis aims to explore and understand the use of orthogonal convolutional neural networks to improve learning of diverse, generic features that are transferable to a novel task.
In this thesis, orthogonal regularization is used to pre-train deep CNNs to investigate if and how orthogonal convolution may improve feature extraction in transfer learning. Experiments using two limited medical image datasets in this thesis suggests that orthogonal regularization improves generality and reduces redundancy of learned features more effectively in certain deep networks for transfer learning. The results on feature selection and classification demonstrate the improvement in transferred features helps select more expressive features that improves generalization performance. To understand the effectiveness of orthogonal regularization on different architectures, this work studies the effects of residual learning on orthogonal convolution. Specifically, this work examines the presence of residual connections and its effects on feature similarities and show residual learning blocks help orthogonal convolution better preserve feature diversity across convolutional layers of a network and alleviate the increase in feature similarities caused by depth, demonstrating the importance of residual learning in making orthogonal convolution more effective.
ContributorsChan, Tsz (Author) / Li, Baoxin (Thesis advisor) / Liang, Jianming (Committee member) / Yang, Yezhou (Committee member) / Arizona State University (Publisher)
Created2023
Description
Vision Transformers (ViT) achieve state-of-the-art performance on image classification tasks. However, their massive size makes them unsuitable for edge devices. Unlike CNNs, limited research has been conducted on the compression of ViTs. This thesis work proposes the ”adjoined training technique” to compress any transformer based architecture. The architecture, Adjoined Vision Transformer (AN-ViT), achieves state-of-the-art performance on the ImageNet classification task. With the base network as Swin Transformer, AN-ViT with 4.1× fewer parameters and 5.5× fewer floating point operations (FLOPs) achieves similar accuracy (within 0.15%). This work further proposes Differentiable Adjoined ViT (DAN-ViT), whichuses neural architecture search to find hyper-parameters of our model. DAN-ViT outperforms the current state-of-the-art methods including Swin-Transformers by about ∼ 0.07% and achieves 85.27% top-1 accuracy on the ImageNet dataset while using 2.2× fewer parameters and with 2.2× fewer FLOPs.
ContributorsGoel, Rajeev (Author) / Yang, Yingzhen (Thesis advisor) / Yang, Yezhou (Committee member) / Zou, Jia (Committee member) / Arizona State University (Publisher)
Created2023