Info iconThis preview shows pages 1–2. Sign up to view the full content.

View Full Document Right Arrow Icon
Convolutional Face Finder: A Neural Architecture for Fast and Robust Face Detection Christophe Garcia and Manolis Delakis Abstract —In this paper, we present a novel face detection approach based on a convolutional neural architecture, designed to robustly detect highly variable face patterns, rotated up to ± 20 degrees in image plane and turned up to ± 60 degrees, in complex real world images. The proposed system automatically synthesizes simple problem-specific feature extractors from a training set of face and nonface patterns, without making any assumptions or using any hand-made design concerning the features to extract or the areas of the face pattern to analyze. The face detection procedure acts like a pipeline of simple convolution and subsampling modules that treat the raw input image as a whole. We therefore show that an efficient face detection system does not require any costly local preprocessing before classification of image areas. The proposed scheme provides very high detection rate with a particularly low level of false positives, demonstrated on difficult test sets, without requiring the use of multiple networks for handling difficult cases. We present extensive experimental results illustrating the efficiency of the proposed approach on difficult test sets and including an in- depth sensitivity analysis with respect to the degrees of variability of the face patterns. Index Terms —Face detection, neural networks, machine learning, convolutional networks. æ 1I NTRODUCTION F ACE detection is becoming a very important research topic, due to its wide range of possible applications, like security access control, model-based video coding, content- based video indexing, or advanced human and computer interaction. It is also a required preliminary step to face recognition and expression analysis. Numerous approaches for face detection have been proposed in the last decade, many of them described and compared in two interesting recent surveys by Yang et al. [1] and Hjelmas et al. [2]. Most face detection methods are based on local facial feature detection and classification using statistical and geometric models of the human face. Low level analysis first deals with the segmentation of visual features using image properties such as edges [3], intensity [4], color [5], [6], motion [7], or generalized measures [8]. Other approaches are based on template matching where several correlation templates are used to detect local subfeatures, considered as rigid in appearance (eigenfeatures [9]) or deformable [10], [11]. Then, visual features are organized into a more global concept of face through facial feature and constellation analysis using face geometry constraints [11], [12], [13], [14]. The main drawback of feature-based approaches is that either little global constraints are applied on the face template or extracted features are significantly influenced by noise, occlusions, and changes in face expression and viewpoint. In order to handle difficult scenarios where multiple faces of
Background image of page 1

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
Image of page 2
This is the end of the preview. Sign up to access the rest of the document.

This note was uploaded on 10/23/2010 for the course COMMINUCAT 123 taught by Professor Ali during the Spring '10 term at Masaryk University.

Page1 / 16


This preview shows document pages 1 - 2. Sign up to view the full document.

View Full Document Right Arrow Icon
Ask a homework question - tutors are online