🌪️IA 100x

What is Artificial Intelligence and how does it work?

An AI solution involves a grouping of various technologies, such as artificial neural networks, algorithms, learning systems, and others that can simulate human capabilities linked to intelligence. For example, reasoning, environmental perception, and the ability to analyze for decision making.

We can say that the concept of AI is related to the ability of technological solutions to perform activities in a way that is considered intelligent.AIs can also "learn for themselves" thanks to learning systems that analyze large volumes of data, enabling them to extend their knowledge.

Artificial Intelligence (AI) is also a field of science whose purpose is to study, develop, and employ machines to perform human activities autonomously.It is also linked to robotics, Machine Learning, voice and vision recognition, among other technologies.

AI Development with First100x 2023

We will use AI protocols for face development and together with the metaverse integrate our investors into digital formats, with that we have some modules that are in development:

Facial Recognition Module - Having the neural network trained, the software can be applied to videos for real-time recognition of people. This module is used in the application of the ML algorithm trained in the recognition of people in real time. Different methods have been Different methods for face recognition have been developed and are being tested. In the tests performed, the methods EigenFaceRecognizer and LBPHFaceRecognizer have presented the best results.

EigenFaceRecognizer method.

// Import Cv2

detectorface = cv2.cascadeclassifier("haarcascade-frontalface-default.xml")
recognizzer = cv2.face.eigenfacerecognezier_create()
recognizzer. read (classificadoreigen.yml')
width, height = 220, 200
font = cv2.FONT_HERSHEY_COMPLEX_SMALL
Camera = cv2.videocapture(1)

LBPHFaceRecognizer method

// Import cv2

detectorface = cv2.cascadeclassifier'(harrcascade-frontalface-default.xml')
recognizzer - cv2.face.LBPHFaceRecognizer_CREATE()
recognizzer. read (classificadorLBPH.yml')
width, height = 220, 200
font = cv2.FONT_HERSHEY_COMPLEX_SMALL
winName = 'TEST WINDOWS for o SOPT'
CV2.NAMEWINDOWPROPERTY (winName, Cv2 WIND_PROP_FULLSCREEN, Cv2. WINDOW_FULLSCREEN)
camera= cv2.VideoCapture.

Conf = 180,0

Since humans have an ability to recognize people's faces, RF systems RF systems try to mimic this ability. That is, images are transmitted to the that the software captures the nodal points, which are around 80, such as the distance between for example the distance between the eyes, the size of the chin, length of the nose, among others.

In our implementation, such information, of faces to be recognized, is analyzed over This means that these nodal points that are to be recognized are analyzed on these characteristics by means of the tested methods. That is, these nodal points that have been collected are measured and stored in a database to form the facial signature, completing the feature collection phase. completing the feature collection phase.

Soon after, in the image recognition phase, which can be exposed in the form of an image or video, the face information is collected, such as proximity and the number of nodal points on each individual's face. nodal points of the face of each individual, such as the distance from the eyes, the length of the nose and mouth and mouth length, and are compared with the information that is already stored in the machine's memory of the machine to search for the owner of the identified face. This image tracking is done by from the capture of the image or video by a computer webcam, cell phone or an image already stored.

Therefore, several problems have been analyzed that can hinder recognition.

Among them are lighting, image resolution, and face movements. But it is still a new a new technology in constant advance, which still presents points to be improved.