General Background: Deep image matting is a fundamental task in computer vision, enabling precise foreground extraction from complex backgrounds, with applications in augmented reality, computer graphics, and video processing. Specific Background: Despite advancements in deep learning-based methods, preserving fine details such as hair and transparency remains a challenge. Knowledge Gap: Existing approaches struggle with accuracy and efficiency, necessitating novel techniques to enhance matting precision. Aims: This study integrates deep learning with fusion techniques to improve alpha matte estimation, proposing a lightweight U-Net model incorporating color-space fusion and preprocessing. Results: Experiments using the AdobeComposition-1k dataset demonstrate superior performance compared to traditional methods, achieving higher accuracy, faster processing speed, and improved boundary preservation. Novelty: The proposed model effectively combines deep learning with fusion techniques, enhancing matting quality while maintaining robustness across various environmental conditions. Implications: These findings highlight the potential of integrating fusion techniques with deep learning for image matting, offering valuable insights for future research in automated image processing applications, including augmented reality, gaming, and interactive video technologies. Highlights: Better Precision: Fusion techniques enhance fine detail preservation. Faster Processing: Lightweight U-Net improves speed and accuracy. Wide Applications: Useful for AR, gaming, and video processing. Keywords: Deep image matting, computer vision, deep learning, fusion techniques, U-Net
Background: The marginal seal is essential for sealant success because penetration of bacteria under the sealant might allow caries onset or progression. The aim of the present study was to estimate and compare the microleakage of pit and fissure sealant after various methods of occlusal surface preparation. Materials and methods: Thirty non-carious premolars extracted for orthodontic reasons were equally divided into three groups. In group one, occlusal fissures were opened with round carbide bur, in group two, occlusal surfaces of the teeth were cleaned with a dry pointed bristle brush and samples of group three were cleaned with a slurry of fine flour of pumice in water using rubber cup. Then fissures of all teeth were etched using 35% p
... Show More
(12)
After the year 2003 terrorist attacks knock Baghdad city capital of Iraq using bomb explosion various, shook the nation, and made public resident of Baghdad aware of the need for better ways to protect occupants, assets, and buildings cause the terrorist gangs adopt style burst of blast to injury vulnerability a wider range form, and many structures will suffer damage from air blast when the overpressure concomitant the blast wave, (i.e., the excess over the atmospheric pressure 14.7 pounds per square inch at standard sea level conditions are about one-half pound per square inch or more(
to attainment injury. Then, the distance to which this overpressure level will extend depends primarily on the energy yield (§1.20) of the burst of
Abstract :In this study, amygdaline in Iraqi plant seeds was extracted and isolated from their seeds matrix using reflux procedure and subsequently identified and determined by high performance liquid chromatography (HPLC) on reversed phase column of LC-18 (150mm x 4.6mm, 5?m )with actonitrile :water ( 50 : 50 ) as mobile phase at flow rate of ( 0.5 mL/min ) and detection at wavelength of 215 nm.The experimental results indicated that the linearity of calibration is in the range of 1.0-30.0 mg L-1amygdaline with the correlation coefficient of 0.9949. The limit of detection (LOD) and limit of quantitation (LOQ) for amygdaline were of 0.88 and 2.93 mg L-1 in standard pure sample. The mean recovery percent is 97.34±0.58 at 95% confidence inte
... Show More
(2)
سها علي حسين, هويدة إسماعيل إبراهيم, Journal of Physical Education, 2017 - Cited by 1
The convergence speed is the most important feature of Back-Propagation (BP) algorithm. A lot of improvements were proposed to this algorithm since its presentation, in order to speed up the convergence phase. In this paper, a new modified BP algorithm called Speeding up Back-Propagation Learning (SUBPL) algorithm is proposed and compared to the standard BP. Different data sets were implemented and experimented to verify the improvement in SUBPL.
The aim of this study was to Identifying The Effect of using Linear programming and Branching programming by computer in Learning and Retention of movement concatenation(Linkwork) in parallel bars in Artistic Gymnastics. The searchers have used the experimental method. The search subject of this article has been taken (30) male - students in the second class from the College of Physical Education/University of Baghdad divided into three groups; the first group applied linear programming by computer, and the second group has been applicated branching programming by computer, while precision group used traditional method in the college. The researchers concluded the results by using the statistical bag for social sciences (spss) such as both
... Show More
(1)
In this study the role of remote sensing and geographic information systems in the planning of some Iraqi cities, including the city of Aziziyah in Kut / Wasit Governorate, was considered. Certain appropriate models (sprawling growth patterns) have been adopted to plan this city, and tables and a future map have been prepared to arrange this city. Land uses for public services will be proposed for adopted case(Aziziyah in Kut / Wasit Governorate). This paper describes a specific application of GIS functionalities for spatial planning analysis, examines the breadth of GIS and attempts to address the analytical part’s limitations in planning studies. The study concluded, through high-resolution visuals, to the possibility of determining th
... Show More