Global ETD Search
Search theses and dissertations gathered from participating repositories worldwide. Every result links back to the library that holds it. No account is needed.
Results
Showing 1 to 20 of 33 for “"Video prediction"”.
-
A Hierarchical Approach to Video Prediction
… learn more about the environment are useful. The prediction and anticipation of future events is a key component of such intelligent decision-making systems. Prediction serves as a useful means to learn useful concepts about the world even from a raw stream of sensory observations, such as images …
-
Video Prediction with Invertible Linear Embeddings
… invertible neural network We predict future video frames from complex dynamic scenes. Our invertible linear embedding (ILE) demonstrates successful learning, prediction and latent state inference. In contrast to other approaches, ILE does not use any explicit reconstruction loss or simplistic …
-
Generative Adversarial Reservoirs for natural video prediction
… can be applied to the problem of natural video prediction. Influenced by previous work, we develop a novel architecture called Generative Adversarial Reservoirs (GAR). We use GARs to predict frames of videos from the UCF-101 dataset and show that although some of the quantitative …
-
Leveraging Task-Conditioned Video Prediction for Robotic Manipulation
… This is performed through a task-conditioned video prediction model and a pose-matching model to match the pose of the robot in the real world robot to the predicted pose from the video prediction model. We first train a vision prediction model that, given a textual description of a …
-
Prior-Driven Spatial-Temporal Learning for Controllable Video Prediction
… that anticipates sensory input and minimizes prediction errors through adaptive feedback. Within computer vision, video prediction has become a central task for investigating this principle, offering a natural setting for spatiotemporal learning. Yet, building an effective world simulator …
-
Explicit object-centric video prediction with deep learning models
Video prediction is a crucial task for intelligent agents such as robots and autonomous vehicles, it enables them to anticipate and act early on time-critical incidents. Many state of-the-art video prediction methods typically model the dynamics of a scene jointly and implicitly and seeing it as a …
-
Im2Vid: Future Video Prediction for Static Image Action Recognition
… capture the action performed. Unlike images, videos have temporal information that greatly improves action recognition by resolving potential ambiguity. We propose to leverage a large amount of readily available unlabeled videos to transfer the temporal information from video domain to static …
-
Geometry-based video prediction with camera motion for mobile robotics
Submission original under an indefinite embargo labeled 'Open Access'. The submission was exported from vireo on 2022-11-11 without embargo terms
-
Advances in deep learning with applications to computer vision and astronomy
… each other. This dissertation look into two video-related tasks in computer vision and propose a novel problem in computational astronomy.</p><p>Specifically, acquiring an in-depth understanding of videos has been a cornerstone problem in computer vision. This problem has been studied by …
-
Advance Video Modeling Techniques for Video Generation and Enhancement Tasks
… advanced techniques that are useful in video modeling for generation and enhancement tasks. In the first part of the thesis, we explore generative modeling that exploits the external corpus for learning priors. The task here is of video prediction, i.e., to extrapolate future sequences …
-
MICROSTRUCTURE EVOLUTION PREDICTION USING DEEP LEARNING
Video prediction models can be used alongside phase field simulations to mitigate some of inherent numerical constraints. A modified version of a simpler yet better video prediction (SimVP) model was used to predict microstructure evolution based on phase field simulations. The model showed good …
-
Learning Perceptual Prediction: Learning from Humans and Reasoning about Objects
… agents acquire sufficient data to train large prediction models? Second, what structure should we embed into these models? In this work, we address both questions. In order to acquire sufficient data, we demonstrate methods for learning prediction models and agent policies from the combination …
-
Learning in networks with communication delays
… learning on Fourier synthesis and autoregressive video prediction tasks.
-
Confidence Distillation for Efficient Action Recognition
… recognizing that they may be wrong about their predictions and measuring the certainty of beliefs, they perform poorly. For one of the most common activation functions, the ReLU and its variants, even a well-calibrated model can produce incorrect but high confidence predictions. In the related …
-
Learning Programs and Algorithms: Toward Efficient and Generalizable AI
… program synthesis, world modeling, code repair, video prediction, and efficient reasoning, we find that this alignment yields order-of-magnitude improvements in sample efficiency, compute efficiency, and generalization that purely statistical scaling does not. ROAP learns programs jointly with …
-
Towards Understanding Robust Neural Coding Through Representational Geometry
… neural network (PredNet) trained for next-frame video prediction, we find that predictive learning yields low-dimensional, low-curvature representations with improved OOD generalization and can also produce brain-like border-ownership signals relevant for scene segmentation. Together, these …
-
Neural Visual Synthesis for Graphics, 4D Gaussians and Videos
… synthesis, including graphics, 4D Gaussians, and videos. First, we propose speeding up Monte Carlo rendering by significantly reducing the number of pixels required for sampling. We introduce a neural network to predict future frames instead of rendering future frames. Second, we introduce a …
-
Stochastic prediction in sequential high-dimensional observation space
… plausible future. This can lead to low-quality predictions in real-world settings with stochastic dynamics. In this thesis, we study the importance of stochasticity in predicting high-quality predictions of the raw sequential observations. We develop a stochastic variational video prediction …
-
Deep Neural Network Design for Video Understanding
This thesis focuses on enhancing video understanding through innovative deep neural network designs, tackling the primary challenges in video analysis tasks. It addresses the critical aspects of handling high dimensionality and data volume, interpreting temporal dynamics and sequential …
-
LEARNING TO DETECT, PREDICT AND SYNTHESIZE OBJECTS
… is challenging. Our work builds on hierarchical video prediction models, which disentangle the video generation into two stages: predicting a high-level representation, such as pose, over time, and then learning a pose-to-pixels translation model for pixel generation. An action sequence for a …
Page 1 of 2