Large scale sign detection using HOG feature variants

Abstract

In this paper we present two variant formulations of the well-known Histogram of Oriented Gradients (HOG) features and provide a comparison of these features on a large scale sign detection problem. The aim of this research is to find features capable of driving further improvements atop a preexisting detection framework used commercially to detect traffic signs on the scale of entire national road networks (1000's of kilometres of video). We assume the computationally efficient framework of a cascade of boosted weak classifiers. Rather than comparing features on the general problem of detection we compare their merits in the final stages of a cascaded detection problem where a feature's ability to reduce error is valued more highly than computational efficiency. Results show the benefit of the two new features on a New Zealand speed sign detection problem. We also note the importance of using non-sign training and validation instances taken from the same video data that contains the training and validation positives. This is attributed to the potential for the more powerful HOG features to overfit on specific local patterns which may be present in alternative video data.

Traffic sign detection and analysis: Recent studies and emerging trends

“…In this paper, we have chosen to cover 4 recent leading papers [7], [9], [10], [11] that describe different methods of detecting signs. These papers, apart from being very recent, cover trends in the area well: Some use theoretical sign models, some use learned models, some are mainly colorbased, some rely more on shapes, some have extensive focus on tracking.…”

“…In [10] they simply opt to not do any segmentation or preprocessing, but jump directly into feature extraction and detection.…”

“…In [10], two different types of Histogram of Oriented Gradient (HOG) features are used. HOG features are, as the name suggests, histograms detailing the orientation of the gradients in an area.…”

“…[11] use a template for where corners should be located. [7], [10] instead use learned classifiers. [7] use a Support Vector Machine (SVM) classifier on the DtB features and [10] use a similar cascaded classifier, trained with LogitBoost.…”

Integrating Deep Learning with Correlation-based Multimedia Semantic Concept Detection

“…It has been drawn researchers' attention that the optimal fusion of multimedia data from different modalities to effectively detect semantic concepts can further benefit other research areas like semantic concept retrieval, surveillance event detection, etc. To overcome the obstacles to multimedia research, some researchers tried to make progress by utilizing highly discriminative and robust features [150] such as Scale Invariable Feature Transformation (SIFT) [94,163] and Histogram of Oriented Gradients (HOG) [152,190]. The idea of considering only a single modality, such as analyzing audio signals for the automatic transcription of speech, leveraging color features for scene recognition, and using temporal features to detect different actions, has also been greatly investigated.…”