scholarly journals A Multi-Level Approach to Waste Object Segmentation

Sensors ◽  
2020 ◽  
Vol 20 (14) ◽  
pp. 3816
Author(s):  
Tao Wang ◽  
Yuanzheng Cai ◽  
Lingyu Liang ◽  
Dongyi Ye
Keyword(s):  
Random Field ◽  
Object Segmentation ◽  
Color Image ◽  
Conditional Random Field ◽  
Depth Image ◽  
Future Research ◽  
Depth Information ◽  
Multi Level ◽  
Multiple Levels ◽  
Potential Object

We address the problem of localizing waste objects from a color image and an optional depth image, which is a key perception component for robotic interaction with such objects. Specifically, our method integrates the intensity and depth information at multiple levels of spatial granularity. Firstly, a scene-level deep network produces an initial coarse segmentation, based on which we select a few potential object regions to zoom in and perform fine segmentation. The results of the above steps are further integrated into a densely connected conditional random field that learns to respect the appearance, depth, and spatial affinities with pixel-level accuracy. In addition, we create a new RGBD waste object segmentation dataset, MJU-Waste, that is made public to facilitate future research in this area. The efficacy of our method is validated on both MJU-Waste and the Trash Annotation in Context (TACO) dataset.

scholarly journals Challenges and Lessons Learned from Multi-Level Multi-Component Interventions to Prevent and Reduce Childhood Obesity

2018 ◽  
Vol 16 (1) ◽  
pp. 30 ◽  
Author(s):  
Joel Gittelsohn ◽  
Rachel Novotny ◽  
Angela Trude ◽  
Jean Butel ◽  
Bent Mikkelsen
Keyword(s):  
Childhood Obesity ◽  
Local Community ◽  
Lessons Learned ◽  
Future Research ◽  
Multiple Components ◽  
Intervention Intensity ◽  
Multi Level ◽  
Multiple Levels ◽  
Study Designs ◽  
Physical Activity Environments

Multi-level multi-component (MLMC) strategies have been recommended to prevent and reduce childhood obesity, but results of such trials have been mixed. The present work discusses lessons learned from three recently completed MLMC interventions to inform future research and policy addressing childhood obesity. B’more Healthy Communities for Kids (BHCK), Children’s Healthy Living (CHL), and Health and Local Community (SoL) trials had distinct cultural contexts, global regions, and study designs, but intervened at multiple levels of the socioecological model with strategies that address multiple components of complex food and physical activity environments to prevent childhood obesity. We discuss four common themes: (i) How to engage with community partners and involve them in development of intervention and study design; (ii) build and maintain intervention intensity by creating mutual promotion and reinforcement of the intervention activities across the multiple levels and components; (iii) conduct process evaluation for monitoring, midcourse corrections, and to engage stakeholder groups; and (iv) sustaining MLMC interventions and its effect by developing enduring and systems focused collaborations. The paper expands on each of these themes with specific lessons learned and presents future directions for MLMC trials.

An automatic technique for presentation of coincident‐loop, impulse‐response, transient, electromagnetic data

Geophysics ◽  
1994 ◽  
Vol 59 (10) ◽  
pp. 1542-1550 ◽  
Author(s):  
Richard S. Smith ◽  
R. N. Edwards ◽  
G. Buselli
Keyword(s):  
Color Image ◽  
Synthetic Data ◽  
Depth Image ◽  
Depth Information ◽  
Color Contrast ◽  
Computationally Efficient ◽  
Apparent Conductivity ◽  
Transient Electromagnetic ◽  
Anomalous Feature ◽  
Depth Curves

Coincident‐loop TEM sounding data are often presented by plotting the half‐space apparent conductivity as a function of delay time. A new algorithm generates an improved presentation that plots the apparent conductivity as a function of depth. The resulting data may be further processed to sharpen or “spike” the smoothly varying apparent‐conductivity/depth curves in an attempt to better represent the rapid changes in conductivity that often exist in the earth. The algorithm described involves an approximation, but is simple, easy to use, and computationally efficient. A layered conductivity structure is assumed, so the algorithm is best for areas where the geology is approximately horizontal. However, the algorithm can also be used to identify anomalous features that are not infinite horizontal layers. The spiked conductivity models derived from synthetic data are consistent with the original layered‐earth models and show a greater resolution than the apparent‐conductivity/depth curves, and sometimes amplify noise in the data. When data are collected along a profile line, the conductivity/depth information can be converted to a color image. For profile data collected over the Elura orebody, the image of the spiked conductivity section shows an anomalous feature at the orebody, and the color contrast is more marked than it is on the apparent‐conductivity/depth image.

scholarly journals A simple modification to the inference process of Murray’s (2020) MIR improves the model’s performance

2021 ◽  
Author(s):  
Yuki Kobayashi
Keyword(s):  
Random Field ◽  
Conditional Random Field ◽  
Future Research ◽  
Research Use ◽  
Inference Process ◽  
Simple Modification ◽  
Input Information ◽  
Observer Model

Murray (2020) recently introduced a novel computational lightness model, Markov Illuminance and Reflectance (MIR), a Bayesian observer model that represents input information and prior assumption with conditional random field (CRF) and that can account for many lightness illusions and phenomena. In the original MIR’s inference process, however, it did not utilize all the links in its CRF. Thus, this letter reports that a simple modification to the original MIR’s inference process improves its performance. MIR is a highly extensible model, so I recommend future research use the proposed version to attain further sophistication.

scholarly journals Multilevel governance: Identity, political contestation, and policy

2020 ◽  
Vol 22 (4) ◽  
pp. 792-799
Author(s):  
Hanna Kleider
Keyword(s):  
European Integration ◽  
Empirical Research ◽  
Multilevel Governance ◽  
Future Research ◽  
Policy Outcomes ◽  
Research Programmes ◽  
Multi Level ◽  
Multiple Levels ◽  
The Eu

This commentary takes stock of how Multi-level Governance and European Integration has helped scholars frame empirical research agendas. It focuses on three specific research programmes emanating from the book: (1) the role of identity in multi-level governance, (2) political contestation in multi-level systems, and (3) the effect of multi-level governance on policy outcomes. It aims to highlight existing knowledge in these lines of research whilst offering several critical reflections and directions for future research. The commentary argues that the book’s observation that governance structures are ultimately shaped by identities rather than by efficiency considerations has proved almost prophetic given recent backlash against the EU. The book expertly shows that there is an inherent tension in sharing authority across multiple levels of government, and that multi-level systems require constant recalibration and renegotiation of how authority is shared.

Sign in / Sign up

Export Citation Format

Share Document