A Neural Network Based Cascaded Classifier for Face Detection in Color Images with Complex Background

Authors:
Kamal Nasrollahi;Mohammad Rahmati;Thomas B. Moeslund
Affiliations:
Laboratory of Computer Vision and Media Technology, Aalborg University, Aalborg øst, Denmark 9220;Pattern Recognition Lab, AmirKabir University of Tech., Tehran, Iran;Laboratory of Computer Vision and Media Technology, Aalborg University, Aalborg øst, Denmark 9220
Venue:
ICIAR '08 Proceedings of the 5th international conference on Image Analysis and Recognition
Year:
2008

Citing 10
Cited 1

Neural Network-Based Face Detection

IEEE Transactions on Pattern Analysis and Machine Intelligence
Face Detection From Color Images Using a Fuzzy Pattern Matching Method

IEEE Transactions on Pattern Analysis and Machine Intelligence
A Fast and Accurate Face Detector Based on Neural Networks

IEEE Transactions on Pattern Analysis and Machine Intelligence
Detecting Faces in Images: A Survey

IEEE Transactions on Pattern Analysis and Machine Intelligence
Convolutional Face Finder: A Neural Architecture for Fast and Robust Face Detection

IEEE Transactions on Pattern Analysis and Machine Intelligence
A robust method for detecting arbitrarily tilted human faces in color images

Pattern Recognition Letters
Face detection using template matching and skin-color information

Neurocomputing
A hierarchical neural network for human face detection

Pattern Recognition
Face recognition/detection by probabilistic decision-based neural network

IEEE Transactions on Neural Networks
Face detection based on skin color in video images with dynamic background

ICIAR'07 Proceedings of the 4th international conference on Image Analysis and Recognition

Face Quality Assessment System in Video Sequences

Biometrics and Identity Management

Quantified Score

Hi-index	0.01

Visualization

Abstract

When utilizing neural networks as a classifier in face detection systems there are two important problems which should be solved: 1. High computations between the network layers and 2. Adjusting the topology of the network. The proposed system in this paper uses a genetic algorithm to directly solve the second problem and a fuzzy inference engine as a pre-classifier to indirectly deal with the first problem. After computing a small number of reliable and easy to extract features from skin like regions, in the pre-classification step, a set of flexible rules are applied by a fuzzy inference engine. The accepted regions are fed into a neural network for final decision making. Using this combination of classifiers has established an acceptable tradeoff between the computation and the missed faces while the rate of correct detection is acceptably high.