Title :
Pedestrian Detection Utilizing Gradient Orientation Histograms and Color Self Similarities Descriptors
Author :
Cosmo, Daniel Luis ; Ottoni Teatini Salles, Evandro ; Marques Ciarelli, Patrick
Author_Institution :
Univ. Fed. do Espirito Santo, Vitoria, Brazil
fDate :
7/1/2015 12:00:00 AM
Abstract :
This paper presents a pedestrian detection system in non-controlled environments based on sliding windows. Systems of this type are based on two major blocks: one for feature extraction and other for window classification. Two techniques for feature extraction are used: HOG (Histogram of Oriented Gradient) and CSS (Color Self Similarities), and to classify windows we use linear SVM (Support Vector Machines). Beyond these techniques, we use mean shift and hierarchical clustering to fuse multiple overlapping detections. To improve the system performance, each descriptor is separately classified using an assemble of SVMs. The results obtained on the dataset INRIA Person show that the proposed system, using only HOG descriptors, achieves better results over similar systems. These results were possible due to the cutting of the final detections to better adapt them to the modified annotations, and some modifications on the parameters of the descriptors. The addition of the modified CSS descriptor to the HOG descriptor increases the efficiency of the system, leading to a log average miss rate equal to 36.2%, when classifying each descriptor separately.
Keywords :
feature extraction; image colour analysis; pedestrians; HOG descriptors; color self similarities descriptors; dataset INRIA Person; feature extraction; gradient orientation histograms; hierarchical clustering; histogram of oriented gradient; linear SVM; mean shift; multiple overlapping detection fusing; noncontrolled environments; pedestrian detection system; sliding windows; support vector machines; window classification; Cascading style sheets; Color; Databases; Feature extraction; Histograms; Image color analysis; Support vector machines; Color Self Similarities; Hierarchical Clustering; Histogram of Oriented Gradient; Mean Shift; Pedestrian Detection;
Journal_Title :
Latin America Transactions, IEEE (Revista IEEE America Latina)
DOI :
10.1109/TLA.2015.7273807