Pedestrian detection in thermal images using adaptive fuzzy C-means clustering and convolutional neural networks

Abstract

Pedestrian detection is paramount for advanced driver assistance systems (ADAS) and autonomous driving. As a key technology in computer vision, it also finds many other applications, such as security and surveillance etc. Generally, pedestrian detection is conducted for images in visible spectrum, which are not suitable for night time detection. Infrared (IR) or thermal imaging is often adopted for night time due to its capability of capturing the emitted energy from pedestrians. The detection process firstly extracts candidate pedestrians from the captured IR image. Robust feature descriptors are formulated to represent those candidates. A binary classification of the extract features is then performed with trained classifier models. In this paper, an algorithm for pedestrian detection from IR image is proposed, where an adaptive fuzzy C-means clustering and convolutional neural networks are adopted. The adaptive fuzzy C-means clustering is used to segment the IR images and retrieve the candidate pedestrians. The candidate pedestrians are then pruned using human posture characteristics and the second central moments ellipse. The convolutional neural network is used to simultaneously learn relevant features and perform the binary classification. The performance of the proposed algorithm is compared with state-of-the-art algorithms on publicly available data set. A better detection accuracy with reduced computational accuracy is achieved.

Autonomous vehicles: challenges, opportunities, and future implications for transportation policies

“…Since their coverage range is limited to close distances, they are viewed capable in detecting lane departures [110]. Furthermore, infrared sensors are used in detecting pedestrians and bicycles, particularly at night [111].…”

Detección de peatones en la noche usando Faster R-CNN e imágenes infrarrojas

“…Así, la detección de puntos de calor en resolución multiespectral usando IFCNN (Illumination Fully Connected Neural Network) ha sido propuesta por Guan et al [8]. Vijay et al [20] añaden una red neuronal convolucional al trabajo de Chen et al [19], para la clasificación. Kim et al [23] han usado cá-maras en el espectro visible para detectar peatones en la noche usando CNN.…”

Context-aware Pedestrian Detection with Salient Region Self-growing in Far-infrared Images

“…In the feature extraction step, both traditional and deep learning features have been used, like histogram of oriented gradients (HOG) [4], local binary pattern (LBP) [5], variations of the two, and convolution features [6]. The extracted features will then be classified as pedestrians or non-pedestrians with a machine learning algorithm, such as support vector machine (SVM) [4], AdaBoost [4], sparse representation classifiers [4] and convolution neural network (CNN) [6].…”

“…The extracted features will then be classified as pedestrians or non-pedestrians with a machine learning algorithm, such as support vector machine (SVM) [4], AdaBoost [4], sparse representation classifiers [4] and convolution neural network (CNN) [6]. CNN and deep learning has proved its superiority over traditional methods on a lot of computer vision problems, but there still exists a kind of hard negative sample in infrared images, which cannot be correctly detected even with a CNN-based classifier.…”