Shape binary patterns: an efficient local descriptor and keypoint detector for point clouds

Many of the research problems in robot vision involve the detection of keypoints, areas with salient information in the input images and the generation of local descriptors, that encode relevant information for such keypoints. Computer vision solutions have recently relied on Deep Learning techniques, which make extensive use of the computational capabilities available. In autonomous robots, these capabilities are usually limited and, consequently, images cannot be processed adequately. For this reason, some robot vision tasks still benefit from a more classic approach based on keypoint detectors and local descriptors. In 2D images, the use of binary representations for visual tasks has shown that, with lower computational requirements, they can obtain a performance comparable to classic real-value techniques. However, these achievements have not been fully translated to 3D images, where research is mainly focused on real-value approaches. Thus, in this paper, we propose a keypoint detector and local descriptor based on 3D binary patterns. The experimentation demonstrates that our proposal is competitive against state-of-the-art techniques, while its processing can be performed more efficiently.

In Silico Pharmacokinetics, ADMET Study and Conceptual DFT Analysis of Two Plant Cyclopeptides Isolated From Rosaceae as a Computational Peptidology Approach

This research presents the outcomes of a computational determination of the chemical reactivity and bioactivity properties of two plant cyclopeptides isolated from Rosaceae through the consideration of Computational Peptidology (CP), a protocol employed previously in the research of similar molecular systems. CP allows the prediction of the global and local descriptors that are the integral foundations of Conceptual Density Functional Theory (CDFT) and which could help in getting in the understanding of the chemical reactivity properties of the two plant cyclopeptides under study, hoping that they could be related to their bioactivity. The methodology based on the Koopmans in DFT (KID) approach and the MN12SX/Def2TZVP/H2O model chemistry has been successfully validated. Various Chemoinformatics tools have been used to improve the process of virtual screening, thus identifying some additional properties of these two plant cyclopeptides connected to their ability to behave as potentially useful drugs. With the further objective of analyzing their bioactivity, the CP protocol is complemented with the estimation of some useful parameters related to pharmacokinetics, their predicted biological targets, and the Absorption, Distribution, Metabolism, Excretion and Toxicity (ADMET) parameters related to the bioavailability of the two plant cyclopeptides under study are also reported.

Analysis of Random Local Descriptors in Face Recognition

This article describes and analyzes the new feature extraction technique, Random Local Descriptor (RLD), that is used for the Permutation Coding Neural Classifier (PCNC), and compares it with Local Binary Pattern (LBP-based) feature extraction. The paper presents a model of face feature detection using local descriptors, and describes an improvement on the PCNC for the recognition of plane rotated and small displaced face images, as applied to three databases, i.e., ORL, FRAV3D and FEI. All databases are described along with the recognition results that were obtained. We also include a comparison of our classifier with the Support Vector Machine (SVM) and Iterative Closest Point (ICP). The ORL database was selected to compare our RLDs with LBP-based algorithms. The PCNC with the RLDs demonstrated the best recognition rate, i.e., 97.49%, in comparison with 90.49% for LBPs. For the FEI image database, we obtained the best recognition rate, i.e., 93.57%, in comparison with 66.74% for LBPs. Using the RLDs and rotating the original images for FRAV3D, we improved the recognition rate by decreasing by approximately twice the number of errors. In addition, we analyzed the influence of different RLD parameters on the quality of facial recognition.

Intelligent System for Parasitized Malaria Infection Detection Using Local Descriptors

Malaria is a curative disease, with therapeutics available for patients, such as drugs that can prevent future malaria infections in countries vulnerable to malaria. Though, there is no effective malaria vaccine until now, although it is an interesting research area in medicine. Local descriptors of blood smear image are exploited in this paper to solve parasitized malaria infection detection problem. Swarm intelligence is used to separate the red blood cells from the background of the blood slide image in adaptive manner. After that, the effective corner points are detected and localized using Harris corner detection method. Two types of local descriptors are generated from the local regions of the effective corners which are Gabor based features and color based features. The extracted features are finally fed to Deep Belief Network (DBN) for classification purpose. Different tests were performed and different combinations of feature types are attempted. The achieved results showed that when using combined vectors of local descriptors, the system gives the desired accuracy which is 100%. The achieved result demonstrates the effectiveness of using local descriptors in solving malaria infection detection problem.