Preferred Language
Articles
/
GheZh5IBVTCNdQwCRLMn
Improving Pre-trained CNN-LSTM Models for Image Captioning with Hyper-Parameter Optimization
...Show More Authors

The issue of image captioning, which comprises automatic text generation to understand an image’s visual information, has become feasible with the developments in object recognition and image classification. Deep learning has received much interest from the scientific community and can be very useful in real-world applications. The proposed image captioning approach involves the use of Convolution Neural Network (CNN) pre-trained models combined with Long Short Term Memory (LSTM) to generate image captions. The process includes two stages. The first stage entails training the CNN-LSTM models using baseline hyper-parameters and the second stage encompasses training CNN-LSTM models by optimizing and adjusting the hyper-parameters of the previous stage. Improvements include the use of a new activation function, regular parameter tuning, and an improved learning rate in the later stages of training. The experimental results on the flickr8k dataset showed a noticeable and satisfactory improvement in the second stage, where a clear increment was achieved in the evaluation metrics Bleu1-4, Meteor, and Rouge-L. This increment confirmed the effectiveness of the alterations and highlighted the importance of hyper-parameter tuning in improving the performance of CNN-LSTM models in image caption tasks.

Scopus Crossref
View Publication
Publication Date
Mon Aug 17 2020
Journal Name
Biochemical And Cellular Archives
SERUM FERRITIN AS A BIOMARKER FOR THYROID GLAND (HYPO AND HYPER) THYRODISM WITH METABOLIC SYNDROME
...Show More Authors

Publication Date
Sat Dec 02 2023
Journal Name
Journal Of Engineering
Performance Assessment of Pile Models Chemically Grouted by Low-Pressure Injection Laboratory Device for Improving Loose Sand
...Show More Authors

The complexity and partially defined nature of jet grouting make it hard to predict the performance of grouted piles. So the trials of cement injection at a location with similar soil properties as the erecting site are necessary to assess the performance of the grouted piles. Nevertheless, instead of executing trial-injected piles at the pilot site, which wastes money, time, and effort, the laboratory cement injection devices are essential alternatives for evaluating soil injection ability. This study assesses the performance of a low-pressure laboratory grouting device by improving loose sandy soil injected using binders formed of Silica Fume (SF) as a chemical admixture (10% of Ordinary Portland Cement OPC mass) to di

... Show More
View Publication Preview PDF
Crossref
Publication Date
Mon Feb 01 2016
Journal Name
Swarm And Evolutionary Computation
Improving the performance of evolutionary multi-objective co-clustering models for community detection in complex social networks
...Show More Authors

Scopus (33)
Crossref (27)
Scopus Clarivate Crossref
Publication Date
Tue Mar 11 2025
Journal Name
Iraqi Journal Of Mechanical And Material Engineering
STUDY ON THE PARAMETER OPTIMIZATION INMAGNETIC ABRASIVE POLISHING FORBRASS CUZN33PLATE USING TAGUCHI METHOD
...Show More Authors

This paper describes a new finishing process using magnetic abrasives were newly made to finish effectively brass plate that is very difficult to be polished by the conventional machining processes. Taguchi experimental design method was adopted for evaluating the effect of the process parameters on the improvement of the surface roughness and hardness by the magnetic abrasive polishing. The process parameters are: the applied current to the inductor, the working gap between the workpiece and the inductor, the rotational speed and the volume of powder. The analysis of variance(ANOVA) was analyzed using statistical software to identify the optimal conditions for better surface roughness and hardness. Regressions models based on statistical m

... Show More
View Publication
Publication Date
Thu Jun 01 2023
Journal Name
Baghdad Science Journal
Comparison of Faster R-CNN and YOLOv5 for Overlapping Objects Recognition
...Show More Authors

Classifying an overlapping object is one of the main challenges faced by researchers who work in object detection and recognition. Most of the available algorithms that have been developed are only able to classify or recognize objects which are either individually separated from each other or a single object in a scene(s), but not overlapping kitchen utensil objects. In this project, Faster R-CNN and YOLOv5 algorithms were proposed to detect and classify an overlapping object in a kitchen area.  The YOLOv5 and Faster R-CNN were applied to overlapping objects where the filter or kernel that are expected to be able to separate the overlapping object in the dedicated layer of applying models. A kitchen utensil benchmark image database and

... Show More
View Publication Preview PDF
Scopus (6)
Crossref (2)
Scopus Crossref
Publication Date
Sat Sep 01 2012
Journal Name
2012 International Conference On Statistics In Science, Business And Engineering (icssbe)
A proposal method for selecting smoothing parameter with missing values
...Show More Authors

In this paper we proposed a new method for selecting a smoothing parameter in kernel estimator to estimate a nonparametric regression function in the presence of missing values. The proposed method is based on work on the golden ratio and Surah AL-E-Imran in the Qur'an. Simulation experiments were conducted to study a small sample behavior. The results proved the superiority the proposed on the competition method for selecting smoothing parameter.

View Publication
Scopus (1)
Scopus Crossref
Publication Date
Sun Jun 01 2014
Journal Name
Journal Of Economics And Administrative Sciences
Different Methods for Estimating Location Parameter & Scale Parameter for Extreme Value Distribution
...Show More Authors

      In this study, different methods were used for estimating location parameter  and scale parameter for extreme value distribution, such as maximum likelihood estimation (MLE) , method of moment  estimation (ME),and approximation  estimators based on percentiles which is called white method in estimation, as the extreme value distribution is one of exponential distributions. Least squares estimation (OLS) was used, weighted least squares estimation (WLS), ridge regression estimation (Rig), and adjusted ridge regression estimation (ARig) were used. Two parameters for expected value to the percentile  as estimation for distribution f

... Show More
View Publication Preview PDF
Crossref
Publication Date
Sat Oct 21 2017
Journal Name
Journal Of New Theory
Cubic hyper KU-ideals
...Show More Authors

It is known that, the concept of hyper KU-algebras is a generalization of KU-algebras. In this paper, we define cubic (strong, weak,s-weak) hyper KU-ideals of hyper KU-algebras and related properties are investigated.

Publication Date
Thu Dec 01 2022
Journal Name
Iraqi Journal Of Science
PLAGIARISM DETECTION SYSTEM IN SCIENTIFIC PUBLICATION USING LSTM NETWORKS
...Show More Authors

Scopus (1)
Scopus
Publication Date
Thu Jun 16 2022
Journal Name
Periodicals Of Engineering And Natural Sciences (pen)
Optimization algorithms for transportation problems with stochastic demand
...Show More Authors

The purpose of this paper is to solve the stochastic demand for the unbalanced transport problem using heuristic algorithms to obtain the optimum solution, by minimizing the costs of transporting the gasoline product for the Oil Products Distribution Company of the Iraqi Ministry of Oil. The most important conclusions that were reached are the results prove the possibility of solving the random transportation problem when the demand is uncertain by the stochastic programming model. The most obvious finding to emerge from this work is that the genetic algorithm was able to address the problems of unbalanced transport, And the possibility of applying the model approved by the oil products distribution company in the Iraqi Ministry of Oil to m

... Show More
View Publication
Scopus (9)
Crossref (3)
Scopus Crossref