Performance improvement of character recognition in industrial applications using prior knowledge for more reliable segmentation

2013 ◽  
Vol 40 (17) ◽  
pp. 6955-6963 ◽  
Author(s):  
Martin Grafmüller ◽  
Jürgen Beyerer
Keyword(s):  
Prior Knowledge ◽  
Performance Improvement ◽  
Character Recognition ◽  
Industrial Applications

Parsing of Urban Facades from 3D Point Clouds Based on a Novel Multi-View Domain

2021 ◽  
Vol 87 (4) ◽  
pp. 283-293
Author(s):  
Wei Wang ◽  
Yuan Xu ◽  
Yingchao Ren ◽  
Gang Wang
Keyword(s):  
Deep Learning ◽  
Prior Knowledge ◽  
Performance Improvement ◽  
Data Distribution ◽  
Point Clouds ◽  
Learning Models ◽  
Data Set ◽  
3D Point Clouds ◽  
Segmentation Accuracy ◽  
The Mean

Recently, performance improvement in facade parsing from 3D point clouds has been brought about by designing more complex network structures, which cost huge computing resources and do not take full advantage of prior knowledge of facade structure. Instead, from the perspective of data distribution, we construct a new hierarchical mesh multi-view data domain based on the characteristics of facade objects to achieve fusion of deep-learning models and prior knowledge, thereby significantly improving segmentation accuracy. We comprehensively evaluate the current mainstream method on the RueMonge 2014 data set and demonstrate the superiority of our method. The mean intersection-over-union index on the facade-parsing task reached 76.41%, which is 2.75% higher than the current best result. In addition, through comparative experiments, the reasons for the performance improvement of the proposed method are further analyzed.

A Computational Fluid Dynamics (CFD) Guided Engineering Solution for Performance Improvement of a Multistage Pump

2021 ◽  
Author(s):  
Teymour Javaherchi ◽  
Susheel Brahmeshwarkar ◽  
Raja Faruq ◽  
Chinmay Deshpande
Keyword(s):  
Performance Improvement ◽  
Test Data ◽  
Industrial Applications ◽  
Pump Efficiency ◽  
Pump Impeller ◽  
Methodology Development ◽  
Engineering Team ◽  
Initial Hypothesis ◽  
Computational Fluid Dynamics Cfd ◽  
Multistage Pump

Abstract This work will demonstrate how the Energy Recovery Inc. (ERI) engineering team improved the efficiency of a multistage pump by about 10% at the first stage, which translated into a 3% increase in the overall multistage pump efficiency; according to a set of engineering calculations and review of the archived in-house test data for the legacy multistage pumps, it was hypothesized that the performance pain-point of the pump was inefficient performance of the first stage, due to the formation of a strong pre-swirl right before its inlet. The validity of this hypothesis then was confirmed via RANS CFD simulations of the flow field inside the inlet suction housing and pump impeller. Same CFD methodology was used to evaluate multiple engineering solutions to reduce the strength of the inflow pre-swirl by modifying the inlet suction housing geometry. The obtained RANS CFD solutions guided the engineering team towards the most promising hardware modification proposal. The proposed geometrical modification of the inlet suction housing was implemented and tested on different multistage pumps. All of the test results validated the obtained RANS CFD numerical solution. The state of the art in this successful performance improvement process was first the on-point hypothesis development based on fundamentals of engineering and archived test data. Second, the proper RANS CFD methodology development to model/confirm the initial hypothesis and vet all possible engineering solutions to maximize the multistage pump efficiently and accurately. This can be a great example for various relevant turbomachinery industrial applications.

An Adaptive PI/Sliding Mode Controller for a Speed Drive

1989 ◽  
Vol 111 (3) ◽  
pp. 409-415 ◽  
Author(s):  
R. M. DeSantis
Keyword(s):  
Performance Improvement ◽  
Sliding Mode ◽  
Experimental Tests ◽  
Industrial Applications ◽  
Open Loop ◽  
Sliding Mode Controller ◽  
Switching Element ◽  
Parallel Addition ◽  
Self Tuning ◽  
Theoretical Results

A classical PI speed drive controller modified with the parallel addition of an on-off switching element appears to offer a potential for reasonable improvement over the performance of the original version. This improvement is obtained by combining classical transfer function techniques, sliding mode systems ideas, and self-tuning. While theoretical results, extended simulations, and preliminary experimental tests are encouraging, they do suggest that in actual industrial applications performance improvement may be conditioned by the usage of better performing open loop components.

scholarly journals On the Performance Improvement of Devanagari Handwritten Character Recognition

2015 ◽  
Vol 2015 ◽  
pp. 1-12 ◽  
Author(s):  
Pratibha Singh ◽  
Ajay Verma ◽  
Narendra S. Chaudhari
Keyword(s):  
Performance Improvement ◽  
Character Recognition ◽  
Gradient Descent ◽  
Stochastic Gradient Descent ◽  
Weight Decay ◽  
Pixel Intensity ◽  
Handwritten Character ◽  
Extraction Algorithm ◽  
Numeral Recognition ◽  
Gradient Feature

The paper is about the application of mini minibatch stochastic gradient descent (SGD) based learning applied to Multilayer Perceptron in the domain of isolated Devanagari handwritten character/numeral recognition. This technique reduces the variance in the estimate of the gradient and often makes better use of the hierarchical memory organization in modern computers.L2-weight decay is added on minibatch SGD to avoid overfitting. The experiments are conducted firstly on the direct pixel intensity values as features. After that, the experiments are performed on the proposed flexible zone based gradient feature extraction algorithm. The results are promising on most of the standard dataset of Devanagari characters/numerals.

scholarly journals Population Based Meta Heuristics Algorithm for Performance Improvement of Feed Forward Neural Network

2020 ◽  
Vol 2 (1) ◽  
pp. 36-46
Author(s):  
Dr. Samuel Manoharan ◽  
Prof. Sathish
Keyword(s):  
Neural Network ◽  
Performance Improvement ◽  
Water Cycle ◽  
Probabilistic Neural Network ◽  
Research Process ◽  
Population Based ◽  
Industrial Applications ◽  
Feed Forward Neural Network ◽  
Feed Forward ◽  
Heuristics Algorithm

The most vital step in mining data’s in order to have a proper decision making is the classification, it is remains important in multiple of human activities such as the industrial applications, marketing campaigns, research process and the scientific endeavors. The process of classifying involves the objects categorization into classes that are already defined. These categorizations are developed according to the identical attributes of the items or the objects. Multitudes of methods were devised to improve the accuracy in the classification to devour an enhanced performance in terms of faster convergence speed. The algorithm based on water cycle that includes the evaporation, condensation and precipitation (WC-ECP), which is a population based metaheuristic is used in the paper to improve the accuracy in the feed forward neural network (PNN-probabilistic neural network) to standardizes its random constraint choice and in turn improvise the accuracy of the categorization and the speed of the convergence. The proposed method was tested with the five dataset of UCI machine learning repository and was evinced that the WCECP-PNN performed better compared to the other evolutionary algorithms such as the GA which is also a population based Meta-heuristics

Possible Approaches for Character Recognition With Existing Methodologies and State-of-the-Art Techniques

2019 ◽  
pp. 232-246
Author(s):  
Rashmi Welekar ◽  
Nileshsingh V. Thakur
Keyword(s):  
Character Recognition ◽  
Optical Character Recognition ◽  
State Of The Art ◽  
Industrial Applications ◽  
New Methods ◽  
Optical Character ◽  
The World ◽  
New Hypothesis ◽  
Key Questions ◽  
Art Techniques

The world started to talk about optical character recognition (OCR) around 1870. Then over another 25 years OCR systems were designed for industrial applications. And now the OCR software is easily available online for free, through products like Acrobat reader, WebOCR, etc. But still the research is on. Do we need to switch direction or introduce new hypothesis are some of the key questions? The purpose of this chapter is to answer the above questions and propose new methods for character recognition.

Sign in / Sign up

Export Citation Format

Share Document