In this paper, we present a novel framework for scene image classification, which depends on corresponding visual words concatenation of speeded up robust features (SURF) and directional binary code (DBC) feature desc...
详细信息
In this paper, we present a novel framework for scene image classification, which depends on corresponding visual words concatenation of speeded up robust features (SURF) and directional binary code (DBC) feature descriptor. Firstly, we use SURF feature descriptor as a local feature descriptor. The local feature descriptor captures very close visual appearance (distinct structure) among their visual contents representation of an image. Secondly, the DBC feature descriptor captures global features, where color-texture features are extracted from entire image. Then, visual words of local and global descriptors are build separately. The concatenated visual words are used to represent the training images and query image. The SVM classifier is used to classify training samples and a query image is classified based on the similarity between histograms of training samples and query image. We carried out experiments using the challenging scene datasets such as MIT scene, UIUC sports event, and MIT indoor scene datasets. The experimental results demonstrate that the proposed method outperforms compared to the existing scene image classification methods.
Face recognition and expression recognition are playing vital roles in various applications such as medical field, entertainment, criminal analysis, social media, online business, etc. Local texture feature descriptor...
详细信息
Face recognition and expression recognition are playing vital roles in various applications such as medical field, entertainment, criminal analysis, social media, online business, etc. Local texture feature descriptors such as LBP, LTrP, LTP, and DBC are usually popular to recognise the faces and expressions as well. In this paper, a new feature descriptor local double directional stride maximum pattern approach is proposed as existing methods are suffering from intensity differences and covering of diagonal information. The proposed approach will identify the facial expression based on the directional coordination of pixels, the pattern will be generated by calculating the first order derivatives in four directions using DBC, and then second order derivatives are calculated maximum and minimum intensity values in four directions among three pixels in every direction to construct the feature. This approach reaches to recognise the major intensity differences by using the directions and pleats the accurate data from an image. Facial expression recognition and retrieval performance is measured and compared in stand of precision, recall and ARR on the benchmark datasets such as JAFFE, CK+, ISED, RaFD, etc. with the existing methods.
Pose and illuminations remain great challenges to current face recognition technique. In this paper, visible image (VI) and near-infrared image (NIR) are fused for performance improvement. When directional binary code...
详细信息
ISBN:
(纸本)9783642259432
Pose and illuminations remain great challenges to current face recognition technique. In this paper, visible image (VI) and near-infrared image (NIR) are fused for performance improvement. When directional binary code is adopted as feature representation. AdaBoost algorithm and the cascade structure are used for classification. Fusion is done at decision level and classification scores are normalized using three different rules, i.e. Min-Max, Z-Score and Tanh-Estimators. Experimental results suggest that the proposed algorithm using VI achieve better performance than NIR when pose and expression variations are present. However, NIR shows much better robustness against illumination and time difference than VI. Due to the complementary information available in two image modalities, fusion of NIR and VI further improves the system performance.
A novel local feature descriptor, namely directional Local binary Patterns (DLBP), was proposed in this paper and applied for face recognition. The descriptor first extracts directional edge information, then codes th...
详细信息
ISBN:
(纸本)9783642254482
A novel local feature descriptor, namely directional Local binary Patterns (DLBP), was proposed in this paper and applied for face recognition. The descriptor first extracts directional edge information, then codes these information using Local binary Patterns (LBP). When applied for face recognition, a face image is divided into a number of small sub-windows, DLBP histogram extracted from each sub-window are then concatenated to form a global description of the face. The proposed method was extensively evaluated on two publicly available databases. i.e. the FERET face database and the PolyU-NIRED near-infrared face database. Experimental results show advantages of DLBP over LBP and directional binary code (DBC).
暂无评论