Global Motion Registration via Long-term Photometric Memory
نویسندگان
چکیده
We introduce an algorithm for causally estimating three-dimensional shape and motion of an object represented locally as a rigid collection of planes supporting projectively deforming texture patches. Due to occlusions and the local nature of any causal algorithm, a drift in the estimates accumulates over time. We describe a method to perform global registration of local estimates of motion and structure by matching the appearance of feature regions stored over long time periods. The irradiance is matched using a score function that takes into account contrast and scaling. We show results on real image sequences that confirm extensive simulation experiments.
منابع مشابه
RGB-D Scene Labeling with Long Short-Term Memorized Fusion Model
Semantic labeling of RGB-D scenes is crucial to many intelligent applications including perceptual robotics. It generates pixelwise and fine-grained label maps from simultaneously sensed photometric (RGB) and depth channels. This paper addresses this problem by i) developing a novel Long Short-Term Memorized Fusion (LSTM-F) Model that captures and fuses contextual information from multiple chan...
متن کاملP7: The Roles of Long-Term Memory on the Organization of the Knowledge for Educators
Modern neuroscientific research help to solve the impotent challenge in curriculum design and teaching for enhancing students’ ability to organize information in a way that makes it efficient in response to an appropriate context such as problem solving and critical thinking via knowing about the mechanism of different type of memories especially long term memory. At first, we should to c...
متن کاملPhotometric Estimation of 3D Surface Motion Fields for Respiration Management
In radiation therapy, the estimation of torso deformations due to respiratory motion is an essential component for real-time tumor tracking solutions. Using range imaging (RI) sensors for continuous monitoring during the treatment, the 3-D surface motion field is reconstructed by a non-rigid registration of the patient’s instantaneous body surface to a reference. Typically, surface registration...
متن کاملThe effect of intrahippocampal microinjection of Naloxone on short –term and long-term memory in adult male rats
Introduction:The hippocampus is one for the major centers of learning and memory. Role of the opioid system has been investigated and on the other hand receptors related to this system such as mu-opioid receptors (MOR) are extended in the hippocampus. In this study the effect of Naloxone administration as a mu opioid receptor antagonist on passive avoidance memory in adult male rats was i...
متن کاملLSTM-CF: Unifying Context Modeling and Fusion with LSTMs for RGB-D Scene Labeling
Semantic labeling of RGB-D scenes is crucial to many intelligent applications including perceptual robotics. It generates pixelwise and fine-grained label maps from simultaneously sensed photometric (RGB) and depth channels. This paper addresses this problem by i) developing a novel Long Short-Term Memorized Context Fusion (LSTMCF) Model that captures and fuses contextual information from multi...
متن کامل