Perbandingan Metode Deteksi Wajah Menggunakan OpenCV Haar Cascade, OpenCV Single Shot Multibox Detector (SSD) dan DLib CNN

Abstract

Comparison of methods in face detection is needed to provide recommendation of best method. This study compared three methods in face detection, namely OpenCV haar cascade, OpenCV Single Shot Multibox Detector (SSD) and Dlib CNN. Face detection is focused on five challenging conditions, namely face detection in head position obstacles, wearing face masks, lighting, background images that have a lot of noise, differences in expression. Data testing is taken randomly on google with reference to one image consisting of more than one detected face with wild condition. The results of the comparative analysis in wild condition show that the OpenCV haar cascade has more weaknesses with a performance percentage of 20% compared OpenCV SSD and Dlib CNN method. Performance results of SSD and Dlib CNN have the same performance in the five conditions tested, which is about 80%.

Algoritma Principal Component Analysis (PCA) dan Metode Bounding Box pada Pengenalan Citra Wajah

“…Beberapa algoritma, teknik maupun metode telah banyak digunakan dalam eksperimen pengenalan wajah seperti algoritma PCA [7], [8], viola jones [9], wavelet [2], CNN [10], [11] , haar cascade [11], [12]. Beberapa contoh tersebut memiliki hasil pengenalan wajah yang sesuai, tingkat akurasi tinggi dan mampu mendeteksi wajah sesuai dengan citra data latih.…”

Face Detection Pada Gambar Dengan Menggunakan Opencv Haar Cascade

“…Metode atau algoritma haar cascade termasuk metode tertua yang paling banyak dipakai sampai sekarang. Meskipun banyak algoritma lain yang berkembang haar cascade dianggap masih relevan dan mampu mendeteksi wajah dengan baik (Lia Farokhah, 2021). Haar cascade adalah salah satu metode face-detection yang cukup tua, namun termasuk algoritma yang masih handal yang pernah ditemukan, sebelum metode deep learning menjadi terkenal (Behera, 2020).…”

Face gesture recognition module optimization of intelligent security system based on Raspberry Pi

“…Face recognition is one of the more popular biometric technologies in recent years, this technology is based on the characteristics of the human face information to carry out identity recognition, the normal use of cameras to capture and collect images and videos containing faces, and automatically achieve detection and recognition in the image to track faces [3][4] . The most used technology today is OpenCV (shown below) [6] , a cross-platform computer vision that can run under multiple operating systems, and face recognition is now a mature technology with several libraries to call upon. The Harr cascade classifier used in this paper utilizes the ResNet pre-training model from the Dlib library in C++ as the underlying language [6][7] and uses the return_128d_features () function proposed by the GitHub user 'coneypo' to extract the image by intercepting the desired image in the video serial port where it is located [7] .…”

“…The most used technology today is OpenCV (shown below) [6] , a cross-platform computer vision that can run under multiple operating systems, and face recognition is now a mature technology with several libraries to call upon. The Harr cascade classifier used in this paper utilizes the ResNet pre-training model from the Dlib library in C++ as the underlying language [6][7] and uses the return_128d_features () function proposed by the GitHub user 'coneypo' to extract the image by intercepting the desired image in the video serial port where it is located [7] . The function to extract the 128D features of the image, and after the identification is completed, the 128D features of the face in the camera are extracted from the video stream introduced by the camera for comparison, and a threshold is set to quantify whether the person in the camera is person_A with the required confidence level under the set environment (a certain Euclidean distance), as shown in Figure 4.…”