Chunlin Ji - Academia.edu (original) (raw)

Papers by Chunlin Ji

DEStech Transactions on Materials Science and Engineering, Jun 23, 2017

Iet Microwaves Antennas & Propagation, 2007

Lecture Notes in Computer Science, 2022

The cooperative manipulators can execute a wide range of tasks, such as carrying large or heavy p... more The cooperative manipulators can execute a wide range of tasks, such as carrying large or heavy payloads, which are difficult for a single manipulator. Dual arm manipulators are in typically operative configuration to mimic human, which are of highly flexibility and dexterity. In this paper, we propose a novel coarse-to-fine deep learning model along with investigating the grasp prior loss based on the well-known antipodal force-closure property. The proposed deep learning model predicts the contact configurations in grasping over-loaded and over-sized objects for dual arm manipulators directly from raw RGB images. We first apply detection network to locate the coarse bounding box of objects, further apply a fine-predicting network on the bounding box clipped images to precisely generate two contact configurations via minimizing regression loss and the antipodal grasp prior loss. Extensive experimental results under dense clutter and occlusion strongly demonstrate the effectiveness ...

2017 International Conference on Security, Pattern Analysis, and Cybernetics (SPAC)

Acoustic emotion recognition has been an active research area. This paper presents a new Chinese ... more Acoustic emotion recognition has been an active research area. This paper presents a new Chinese corpus of emotionally colored conversations. Two discrete 3-point scaled emotion primitives are used to describe emotions, namely valence and arousal. Acoustic feature extraction is carried out using OpenSMILE toolkit. For the estimation of these primitives, Support Vector Machine (SVM) is used for the classification task. Preliminary classification results show the effectiveness of the proposed method.

2018 21st International Conference on Information Fusion (FUSION), 2018

Continuous-time autoregressive (CAR) model is very powerful when modeling many real world continu... more Continuous-time autoregressive (CAR) model is very powerful when modeling many real world continuous processes. When the model is driven by Brownian motion, parameter inference is usually based on the likelihood calculation using the Kalman filter; while the model is driven by non-Gaussian Lévy process, Monte Carlo type of methods are often applied to approximate the likelihood. In both cases, likelihood evaluation is the key but is not always easy. Here we propose an innovative Bayesian inference method without the requirement of likelihood evaluation. The algorithm is in a framework of approximate Bayesian computation (ABC). Distance correlation is employed as a very flexible summary statistics for ABC and the p-value calculated from distance correlation provides a good measurement of the dependence between generated samples. Simulation study shows that this approach is straightforward and effective in inferring CAR model parameters.

WIREs Computational Statistics, 2021

Text detection and recognition, which is also known as optical character recognition (OCR), is an... more Text detection and recognition, which is also known as optical character recognition (OCR), is an active research area under quick development with a lot of exciting applications. Deep‐learning‐based methods represent the state‐of‐art of this area. However, these methods are largely deterministic: they give a deterministic output for each input. For both statisticians and general users, methods supporting uncertainty inference are of great appeal, leaving rich research opportunities to incorporate statistical models and methods with the established deep‐learning‐based approaches. In this paper, we provide a comprehensive review of the evolution history of research development on OCR with discussions on the statistical insights behind these developments and potential directions to enhance the current methods with statistical approaches. We hope this article can serve as a useful guidebook for statisticians who are seeking for a path toward edge‐cutting research in this exciting area.

IEICE Transactions on Fundamentals of Electronics, Communications and Computer Sciences, 2018

We discuss novel approaches to evaluation of both upper and lower bounds on log marginal likeliho... more We discuss novel approaches to evaluation of both upper and lower bounds on log marginal likelihoods for model comparison in Bayesian analysis. From posterior Monte Carlo samples, we show how existing variational approximation methods defining lower bounds on marginal likelihoods can be extended to also define upper bounds, and develop optimization methods to minimize such upper bounds. Further, using this new approach to upper bound evaluation, we suggest and exemplify a new quasi-optimized lower bound that can often be obtained with trivial computations compared to current methods. We further discuss the use of partial analytic marginalization of some model parameters as a way of significantly reducing the differences between upper and lower bounds to improve marginal likelihood approximation. To implement this, however, traditional variational methods are intractable, and we provide solution in terms of a novel Monte Carlo Stochastic Approximation (MCSA). We provide theoretical results on convergence of the resulting approximations to true bounds, and several simulation examples in regression and mixture models to demonstrate the accuracy and efficacy ofthe new

2016 IEEE International Conference on Signal Processing, Communications and Computing (ICSPCC), 2016

The design of the communication system follows the Nyquist intersymbol interference (ISI) criteri... more The design of the communication system follows the Nyquist intersymbol interference (ISI) criterion, which explains the design principle of the transmitting and receiving filters to mitigate the effects of ISI. In this work, we study how to utilize the ISI to benefit a communication system. We propose a novel modulation framework, which transmits the constellation symbols much faster than the Nyquist rate and therefore introduces heavy interference between consecutive symbols. The benefit of such modulation is that the frequency bandwidth of the modulated symbols is significantly compressed compared to the traditional modulated symbol with the same symbol rate. We discuss the demodulation process and present a particle filtering-type approximation demodulation method. Numerical simulation and comparison study demonstrate that under an achievable signal noise ration condition the meta-modulation realize a high spectrum efficiency, which is impossible to implement in the traditional QAM system.

Progress In Electromagnetics Research Letters, 2019

The analysis and equivalent circuit modeling of square patch FSS adopting an efficient vector-fit... more The analysis and equivalent circuit modeling of square patch FSS adopting an efficient vector-fitting is proposed. The simulations of microstructure are performed with CST Microwave Studio on single-substrate for different physical parameters, angle of incidences and polarizations. Then circuit model is extracted and developed using vector-fitting tool and implemented in SPICE (Simulation Program with Integrated Circuit Emphasis) simulator for both time and frequency analyses. ADS SPICE generator is used for validating the proposed circuit model. The developed model is within 1% of average deviation against the results obtained from the EM simulation and traditional circuit model.

L'invention concerne un procede de selection d'antennes (6, 10; 32, 48-50) en vue d'u... more L'invention concerne un procede de selection d'antennes (6, 10; 32, 48-50) en vue d'une communication entre le cote recepteur et le cote emetteur d'un systeme MIMO sans fil, le procede comprenant les etapes consistant a obtenir des informations d'etat de canal representatives de fonctions de transfert par canal entre des combinaisons d'antennes emettrices et receptrices, et a executer au moins une fois un algorithme iteratif destine a calculer un vecteur de parametres optimises pour une famille de fonctions parametrees discretes de densite de probabilite attribuant une plus forte probabilite aux etats representant une selection d'antennes donnant une valeur plus souhaitable d'une fonction cible sur la base des informations d'etat de canal et d'une valeur optimisee de la fonction cible calculee pour au moins un etat echantillon pris au moins en partie suivant la fonction de densite de probabilite obtenue a l'aide du vecteur de parametres co...