Face Detection using Haar Lick Features

Face detection plays a significant role in many applications, such as video surveillance, gender classification, and facial recognition. In this paper, we propose a new face detection method based on multi-scale histograms. The proposed method uses a multi-scale histogram to represent a face, thereby improving computational efficiency, and making the process suitable for big-data multimedia databases. While the majority of the existing methods concentrate on indexing high dimensional visual features and also have limitations of scalability, within this paper we advise a scalable method for content-based image retrieval in peer-to-peer systems by using the bag-of-visual words model. The codebook such an atmosphere must be updated periodically, instead of stored static. Within this paper, we present a singular approach to dynamically generate increase a worldwide codebook, which views both discriminability and workload balance. Additionally, a peer-to-peer network frequently evolves dynamically, making a static codebook less efficient for retrieval tasks. To be able to further improve retrieval performance and lower network cost, indexing pruning techniques are developed. In contrast to centralized environments, the important thing challenge would be to efficiently get you a global codebook, as images are distributed over the whole peer-to-peer network.