In the era 2000s, image-based technology evolved so rapidly along with technological advances. One application in the field of face detection research. Research on face detection was first introduced by Viola and Jones researchers in 2001. In addition, this research is motivated by the presence of student attendance on campus which is still manual and not a few students who cheated when present. The topic of this research is optimization of face detection based on image processing so as to get the right technique / method in detecting face image and it can reduce false positive error for non-face object in the classroom. This research was conducted in campus IST Akprind Yogyakarta with the aim of applying automatic presences for student attendance. The methods proposed in this study include the Viola-Jones method for facial detection, feature extraction using 12 color statistics features, and classification process using the Multi Layer Perceptron classifier to optimize the detection process. By using 309 data of face candidates, this research was able to detect face object with accuracy value of 82%, specificity value of 35%, and sensitivity value of 97%. This is shows that the addition of 12 color statistic feature extraction and Multi Layer Perceptron can increase the accuracy value of 6% and the spesificity value of 11%.
Pedestrians are one of the street users who have the right to get priority on security. Highway users such as vehicle drivers sometimes violate the traffic lights that is endanger pedestrians and make pedestrians feel insecure when crossing the street. Based on this problem, a tool is designed to provide a warning for the drivers or riders violating the traffic lights and prevent traffic accident by spraying water. The system is able to detect traffic violation based on changes in the value of the vehicle position on the stop line obtained from the Ultrasonic HC-SR04 sensor. When a violation is detected, a decision tree algorithm turns on the pump to spray water to the traffic violators as a deterrent effect. The results show that the vehicle located closest to the sensor has 94% precision, 88% recall and 85% accuracy, the vehicle located in the middle has 73% precision, 100% recall, and 75% accuracy, and the vehicle located furthest to the sensor has 75% precision, 100% recall and 80% accuracy.
Currently, the rapid development of information technology causes various positive impacts and negative impacts. All information (positive and negative contents) are available on the internet. They can easily accessible by various community members including students. Negative content or pornography contained in the internet can have adverse effects, affect the psychological and mental state, especially among students. The purpose of this research is to develop a system for identifying negative content based on the detection of the body's vital signs. The object of the body's vital signs is the nipple. The proposed method is a combination of face detection and face replace to reduce false positive error in the face area. Furthermore, Haar-Cascade Classifier training uses 1000 positive images data (nipple images) and 8000 negative images data (images that not contain of nipple). The feature extraction stage uses the Gray Level Co-occurrence Matrix (GLCM) 84 attribute and the result is continued YCbCr color space feature extraction process. The classification process use Multi Layer Perceptron with architecture of 10 neurons and 1 hidden layer. By using 158 data of nipple candidate objects, this research was able to detect nipple content with accuracy value of 90,3%, specificity value of 84,60%, and sensitivity value of 92,4%. This is shows that the addition of YCbCr color space feature extraction can increase the accuracy value of 0.9% and the sensitivity value of 1.04%.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.