Histogram of oriented gradients

About: Histogram of oriented gradients is a research topic. Over the lifetime, 2037 publications have been published within this topic receiving 55881 citations. The topic is also known as: HOG.

Curvature enhanced bearing fault diagnosis method using 2D vibration signal

Abstract: As a novel representation method, two dimensional (2D) segmentation is gaining ground as an effective condition monitoring method due to its high-level information descriptional ability. However, the accuracy of extracting frequency information is still limited by the finite gray-level and the extraction ability of distinguishable texture for each fault. To overcome these drawbacks, this research proposes a bearing fault diagnosis method using the converted 2D vibrational signal matrices. In this method, 1D vibration signals are converted into 2D matrices to exploit the fault signatures from the converted images. Curvature filtering (mean curvature) algorithm is applied to eliminate the overwhelming interfering contents and preserves the necessary edge information contained in the 2D matrix. In addition, the histogram of oriented gradients features is employed for the effective fault feature extraction. Finally, a one-versus-one support vector machine is utilized for the automatically fault classification. An experimental investigation was carried out for the performance evaluation of the proposed method. Comparison results indicate that the established method is capable of bearing fault detection with considerable accuracy.

Compressing Feature Sets with Digital Search Trees

Abstract: State-of-the-art image retrieval pipelines are based on “bag-of-words” matching. We note that the original order in which features are extracted from the image is discarded in the “bag-of-words” matching pipeline. As a result, a set of features extracted from a query image can be transmitted in any order. A set of m unique features has m! orderings, and if the order of transmission can be discarded, one can reduce the query size by an additional log 2 (m!) bits. We propose a coding scheme based on Digital Search Trees that reduces size of a set of features by approximately log 2 (m!) bits. We perform analysis of the scheme, and show how it applies to any set of symbols in which order can be discarded. We illustrate how the scheme can be applied to a set of low bitrate Compressed Histogram of Gradients (CHoG) descriptors.

Text detection and recognition in natural scene images

Abstract: Text detection in natural scenes is an important but challenging problem because of variations in the text fonts, size, line orientation, complex background in image and non-uniform illuminations. To overcome these problems, effective features for text image recognition are used.In this paper, a text region detector is designed by using a widely used feature descriptor, histogram of oriented gradients (HOG). Local binarization is applied to segment connected components. For text extraction, parameters like normalized height width ratio and compactness are taken into account to filter out text and non-text components. Text recognition is implemented using zone centroid and image centroid based distance metric feature extraction system.

Machine Cognition of Violence in Videos Using Novel Outlier-Resistant VLAD

Abstract: Understanding highly accurate and real-time violent actions from surveillance videos is a demanding challenge. Our primary contribution of this work is divided into two parts. Firstly, we propose a computationally efficient Bag-of-Words (BoW) pipeline along with improved accuracy of violent videos classification. The novel pipeline's feature extraction stage is implemented with densely sampled Histogram of Oriented Gradients (HOG) and Histogram of Optical Flow (HOF) descriptors rather than Space-Time Interest Point (STIP) based extraction. Secondly, in encoding stage, we propose Outlier-Resistant VLAD (OR-VLAD), a novel higher order statistics-based feature encoding, to improve the original VLAD performance. In classification, efficient Linear Support Vector Machine (LSVM) is employed. The performance of the proposed pipeline is evaluated with three popular violent action datasets. On comparison, our pipeline achieved near perfect classification accuracies over three standard video datasets, outperforming most state-of-the-art approaches and having very low number of vocabulary size compared to previous BoW Models.

Classification of Building Information Model (BIM) Structures with Deep Learning

Abstract: In this work we study an application of machine learning to the construction industry and we use classical and modern machine learning methods to categorize images of building designs into three classes: Apartment building, Industrial building or Other. No real images are used, but only images extracted from Building Information Model (BIM) software, as these are used by the construction industry to store building designs. For this task, we compared four different methods: the first is based on classical machine learning, where Histogram of Oriented Gradients (HOG) was used for feature extraction and a Support Vector Machine (SVM) for classification; the other three methods are based on deep learning, covering common pre-trained networks as well as one designed from scratch. To validate the accuracy of the models, a database of 240 images was used. The accuracy achieved is 57% for the HOG + SVM model, and above 89% for the neural networks.