Face Counting Based on Pre-trained Machine Learning Models: A Brief Systematic Review
摘要
Face detection is an important research topic in computer vision and pattern recognition. It is a challenging topic due to the large degree of variation in the appearance of human faces. This field has made tremendous progress and it is difficult to decide which model to choose for which kind of datasets. In this paper, a comparative study of well-established face detection algorithms such as Multi-task Cascaded Convolutional Neural Networks (MTCNN), RetinaFace (a single-shot multi-level face localisation) and Haar Cascade classifier are performed on our own KLETECH dataset, which comprises varied classes of images. This study also presents a brief description about prior-mentioned algorithms. According to our experiments, RetinaFace shows higher performance as compared to other two models in terms of classification accuracy.