Search

Accurate gathering of phenotypic information is a key aspect in several subject matters, including biometrics, biomedical analysis, forensics, and many other. Automatic identification of anatomical structures of biometric interest, such as fingerprints, iris patterns, or facial traits, are extensively used in applications like access control and anthropological research, all having in common the drawback of requiring intrusive means for acquiring the required information. In this regard, the ear structure has multiple advantages. Not only the ear's biometric markers can be easily captured from the distance with non intrusive methods, but also they experiment almost no changes over time, and are not influenced by facial expressions. Here we present a new method based on Geometric Morphometrics and Deep Learning for automatic ear detection and feature extraction in the form of landmarks. A convolutional neural network was trained with a set of manually landmarked examples. The network is able to provide morphometric landmarks on ears' images automatically, with a performance that matches human landmarking. The feasibility of using ear landmarks as feature vectors opens a novel spectrum of biometrics applications. [ABSTRACT FROM AUTHOR]

Gathering precise and detailed geomorphological and dynamic information of coastal processes is increasingly required for environmental studies and coastal management policies as well. Traditional methods for in situ measurements, or remote sensing monitoring by satellites or airbone imagery, impose limitations and tradeoffs between image quality, operational costs, availability, and negative environmental effects. These limitations and tradeoffs restrict the kind of environmental studies that can be undertaken, specifically when a high spatial and temporal resolution is required over wide geographical areas. In the last decades, video monitoring systems have demonstrated to be a cost-effective alternative for this and other similar purposes. Notwithstanding that, video processing is not fully mature in the context of environmental monitoring in general, and, thus, most of the past and current efforts have been developed in an ad hoc basis. This has the drawback that most available solutions are hardly useful in contexts different from their original setup. In this work we develop an autonomous application for geographic feature extraction and recognition in coastal videos. Specifically, we address the classification and feature measurement of multiple beach zones, a topic addressed to a lesser extent by other projects. The system is designed to be deployed in inexpensive, off-the-shelf hardware, and open source software development frameworks, in a way such that the results can be easily replicated by other research groups. The initial setup and calibration requires very simple supervision, thus allowing the system to be used in a variety of coastal environments. [ABSTRACT FROM AUTHOR]

Coastal environments are spaces where people may develop varied economic and recreational activities, such as tourism, which usually damage beaches and other natural resources commonly placed in these settings. The aim of this paper is to present a methodology to estimate and evaluate the Beach Carrying Capacity (BCC) and the actual beach usage level in coastal cities, using on-site information and video processing to provide significant real-time data. To test our methodology, we chose the coastal city of Monte Hermoso, Argentina, as a pilot site because it is by far the prime choice for a large population during summer vacation in this country. Initially, to estimate BCC, cartographic information about facilities and beach zones was collected and combined with surveys requested to tourists, to better understand their habits and preferences. This allowed an accurate estimation of other beach capacities related to BCC. Secondly, beach video sequences were processed with an algorithm that identified, located and counted people on the beach with an adequate accuracy. The actual occupancy factor was computed and used to asses whether the BCC had been exceeded. Also, people were tracked and their preferred relaxing areas were registered (e.g, closer to seaside during the morning). Finally, all the information was stored and visualized using a Geographic Information System (GIS) which allows both to analyze the different information layers and to produce interactive thematic maps. In this way, the resulting methodology may help to identify zones under risk of deterioration and to define suitable places for the development of varied activities (specially those related to tourism). It may also serve as a dashboard for decision and policy making and contribute to coastal management planning as well. [ABSTRACT FROM AUTHOR]

[Display omitted] • We propose a residual neural network for extracting automatically learned features without prior knowledge. • The method is aimed to automatically classify Iberian ceramic pottery. • We used transfer learning techniques to retrain a residual neural network with profile images tagged by an experts group. • We discuss the relevance of introspection in automatic feature extraction, and the effects of poor feature selection. • Trained models with transfer learning generate more accurate feature representations, and require less amount of data. In Cultural Heritage inquiries, a common requirement is to establish time-based trends between archaeological artifacts belonging to different periods of a given culture, enabling among other things to determine chronological inferences with higher accuracy and precision. Among these, pottery vessels are significantly useful, given their relative abundance in most archaeological sites. However, this very abundance makes difficult and complex an accurate representation, since no two of these vessels are identical, and therefore classification criteria must be justified and applied. For this purpose, we propose the use of deep learning architectures to extract automatically learned features without prior knowledge or engineered features. By means of transfer learning, we retrained a Residual Neural Network with a binary image database of Iberian wheel-made pottery vessels' profiles. These vessels pertain to archaeological sites located in the upper valley of the Guadalquivir River (Spain). The resulting model can provide an accurate feature representation space, which can automatically classify profile images, achieving a mean accuracy of 0.96 with an f -measure of 0.96. This accuracy is remarkably higher than other state-of-the-art machine learning approaches, where several feature extraction techniques were applied together with multiple classifier models. These results provide novel strategies to current research in automatic feature representation and classification of different objects of study within the Archaeology domain. [ABSTRACT FROM AUTHOR]

Accurate classification of pottery vessels is a key aspect in several archaeological inquiries, including documentation of changes in style and ornaments, inference of chronological and ethnic groups, trading routes analyses, and many other matters. We present an unsupervised method for automatic feature extraction and classification of wheel-made vessels. A convolutional neural network was trained with a profile image database from Iberian wheel made pottery vessels found in the upper valley of the Guadalquivir River (Spain). During the design of the model, data augmentation and regularization techniques were implemented to obtain better generalization outcomes. The resulting model is able to provide classification on profile images automatically, with an accuracy mean score of 0.9013. Such computation methods will enhance and complement research on characterization and classification of pottery assemblages based on fragments. [ABSTRACT FROM AUTHOR]