This paper presents an approach for the detection and localization of abnormal events in pedestrian areas. The goal is to design a model to detect abnormal events in video sequences using motion and appearance information. Motion information is represented through the use of the velocity and acceleration of optical flow and the appearance information is represented by texture and optical flow gradient. Unlike literature methods, our proposed approach provides a general solution to detect both global and local abnormal events. Furthermore, in the detection stage, we propose a classification by local regions. Experimental results on UMN and UCSD datasets confirm that the detection accuracy of our method is comparable to state-of-the-art methods.

Worldwide, chronic pain has established as one of the foremost medical issues due to its 35% of comorbidity with depression and many other psychological problems. Traditionally, self-report (VAS scale) or physicist inspection (OPI scale) perform the pain assessment; nonetheless, both methods do not usually coincide [14]. Regarding self-assessment, several patients are not able to complete it objectively, like young children or patients with limited expression abilities. The lack of objectivity in the metrics draws the main problem of the clinical analysis of pain. In response, various efforts have tried concerning the inclusion of objective metrics, among which stand out the Prkachin and Solomon Pain Intensity (PSPI) metric defined by face appearance [5]. This work presents an in-depth learning approach to pain recognition considering deep facial representations and sequence analysis. Contrasting current state-of-the-art deep learning techniques, we correct rigid deformations caught since registration. A preprocessing stage is applied, which includes facial frontalization to untangle facial representations from non-affine transformations, perspective deformations, and outside noises passed since registration. After dealing with unbalanced data, we fine-tune a CNN from a pre-trained model to extract facial features, and then a multilayer RNN exploits temporal relation between video frames. As a result, we overcome state-of-the-art in terms of average accuracy at frames level (80.44%) and sequence level (84.54%) in the UNBC-McMaster Shoulder Pain Expression Archive Database.

Object recognition in the field of computer vision is a constant challenge to achieve better precision in less time. In this research, is proposed a new 3D descriptor, to work with depth cameras called CPPFL, based on the PPF descriptor from [1]. This proposed descriptor takes advantage of color information and groups it more effectively and lightly than the CPPF descriptor from [2], which uses the color information too. Also, it is proposed as an alternative descriptor called CPPFL+, which differs in the construction taking advantage of the same concept of grouping colors, so it gains a “plus” in speed. This change makes the descriptor more efficient compared to PPF and CPPF descriptors. Optimizing the object recognition process [3], it can reach a rate of ten frames per second or more depending on the size of the object. The proposed descriptor is more tolerant to illuminations changes since the hue of a color is more relevant than the other components.

Currently, IoT has evolved to such an extent to extend to all corners of each place through devices that are connected to a network and generate information. In most cases, to be processed for a specific purpose or storage as historical data, an IoT ecosystem is implemented to manage those tasks between different devices, frameworks, or applications. Besides, more complex IoT ecosystems require more complex architecture to manage the information flow; at this level, we found a problem called interoperability. This problem is not limited to the compatibility of adding/removing devices to an ecosystem; it is also expected that the information generated by devices and processed complies with a standard optimizing the data transmission. In this work, we present a new software architecture pattern to avoid the problem of interoperability through the process of exchange information between devices and prevent and store heterogeneous information coming from different layers of an IoT ecosystem.