Deepfake is a type of artificial intelligence used to create convincing images, audio, and video hoaxes and it concerns celebrities and everyone because they are easy to manufacture. Deepfake are hard to recognize by people and current approaches, especially high-quality ones. As a defense against Deepfake techniques, various methods to detect Deepfake in images have been suggested. Most of them had limitations, like only working with one face in an image. The face has to be facing forward, with both eyes and the mouth open, depending on what part of the face they worked on. Other than that, a few focus on the impact of pre-processing steps on the detection accuracy of the models. This paper introduces a framework design focused on this aspect of the Deepfake detection task and proposes pre-processing steps to improve accuracy and close the gap between training and validation results with simple operations. Additionally, it differed from others by dealing with the positions of the face in various directions within the image, distinguishing the concerned face in an image containing multiple faces, and segmentation the face using facial landmarks points. All these were done using face detection, face box attributes, facial landmarks, and key points from the MediaPipe tool with the pre-trained model (DenseNet121). Lastly, the proposed model was evaluated using Deepfake Detection Challenge datasets, and after training for a few epochs, it achieved an accuracy of 97% in detecting the Deepfake
Abstract: Data mining is become very important at the present time, especially with the increase in the area of information it's became huge, so it was necessary to use data mining to contain them and using them, one of the data mining techniques are association rules here using the Pattern Growth method kind enhancer for the apriori. The pattern growth method depends on fp-tree structure, this paper presents modify of fp-tree algorithm called HFMFFP-Growth by divided dataset and for each part take most frequent item in fp-tree so final nodes for conditional tree less than the original fp-tree. And less memory space and time.
Roof in the Iraqi houses normally flattening by a concrete panel. This concrete panel has poor thermal properties. The usage of materials with low thermal conductivity and high specific heat gives a good improvements to the thermal properties of the concrete panel, thus, the indoor room temperature improves. A Mathcad program based on a mathematical model employing complex Fourier series built for a single room building. The model input data are the ambient temperature, solar radiation, and sol-air temperature, which have been treated as a periodic function of time. While, the room construction is constant due to their materials made of it, except the roof properties are taken as a variable generated practically from the
... Show MoreBusiness incubator is a new effective mechanism in developing small projects through its introductions of a new integral system of services. It aims at supporting and developing making new projects. Hence, there is a big number of factors that are interrelated in the processes of preparation for those projects. Those factors are: organizing the incubator and the market available for the projects attached to them and the work programs which will have to be implemented. Those small projects represent more than 98% of the total work institutions in the world. Also it has become responsible for a ration reaching half of the national output of those countries. These projects have created between 40 to 80% of job opportunities availabl
... Show MoreFabrication of a photodetector consists of the conjugated polymer "MEH-PPV"- poly (2-methoxy-5-(2'-ethylhexyloxy)-1,4-phenlenevinylene) and MEH-PPV:MWCNT nanocomposite thin film. The volume ratio investigated was 0.75:0.25. MEH-PPV was dissolved in chloroform solvent and doped with MWCNTs. The spin coating method was used to achieve a facile and low cost photodetector. The absorption spectrum decreases by adding the CNTs. The PL spectrum detected recombination curve results by doping the polymer with CNTs, and AFM measurement showed an increase of roughness average from (0.168 to 2.43nm) of "MEH-PPV" and "MEH-PPV:CNTs", respectively. The doping ratio 0.25, which has a higher photoresponsivity, was evaluated at 1.70 A/W and 2.14 A/W of th
... Show MoreDegenerate parabolic partial differential equations (PDEs) with vanishing or unbounded leading coefficient make the PDE non-uniformly parabolic, and new theories need to be developed in the context of practical applications of such rather unstudied mathematical models arising in porous media, population dynamics, financial mathematics, etc. With this new challenge in mind, this paper considers investigating newly formulated direct and inverse problems associated with non-uniform parabolic PDEs where the leading space- and time-dependent coefficient is allowed to vanish on a non-empty, but zero measure, kernel set. In the context of inverse analysis, we consider the linear but ill-pose
A factor group is a mathematical group obtained by aggregating similar elements of a larger group using an equivalence relation that preserves some of the group structure. In this paper, the factor groups K(SL(2,121)) and K(SL(2,169)) computed for each group from the character table of rational representations.