Human violence recognition is an area of great interest in the scientific community due to its broad spectrum of applications, especially in video surveillance systems, because detecting violence in real time can prevent criminal acts and save lives. The majority of existing proposals and studies focus on result precision, neglecting efficiency and practical implementations. Thus, in this work, we propose a model that is effective and efficient in recognizing human violence in real time. The proposed model consists of three modules: the Spatial Motion Extractor (SME) module, which extracts regions of interest from a frame; the Short Temporal Extractor (STE) module, which extracts temporal characteristics of rapid movements; and the Global Temporal Extractor (GTE) module, which is responsible for identifying long-lasting temporal features and fine-tuning the model. The proposal was evaluated for its efficiency, effectiveness, and ability to operate in real time. The results obtained on the Hockey, Movies, and RWF-2000 datasets demonstrated that this approach is highly efficient compared to various alternatives. In addition, the VioPeru dataset was created, which contains violent and non-violent videos captured by real video surveillance cameras in Peru, to validate the real-time applicability of the model. When tested on this dataset, the effectiveness of our model was superior to the best existing models.

In this work we aim to detect faces in violence scenes, in order to help the security control. We used the Violent Flow (ViF) descriptor with Horn-Schunck proposed in V. Machaca Arceda and K. Fernańdez Fabián and J.C. Gutiérrez, Real Time Violence Detection in Video, \"IET Conference Proceedings\", Institution of Engineering and Technology. (2016) for violence scenes detection at first stage. Then we applied the non-adaptive interpolation super resolution algorithm to improve the video quality and finally we fire a Kanade-Lucas-Tomasi (KLT) face detector. In order to get a very low time processing, we paralleled the super resolution and face detector algorithms with CUDA. For the experiments we used the Boss Dataset and also we built a violence dataset, taking scenes from surveillance cameras. We have promising results detecting faces in this environment, because of the benefits of our proposal.

Evaluating the physical defects of green coffee beans are an important process in defining their quality. This evaluation is normally carried out by visual inspection or using traditional instruments which have some limitations. This work is focused on the implementation of a computer vision system combining a hardware prototype and a software module. The hardware was developed to capture the images of coffee beans, the software uses a White-Patch algorithm as a image enhancement procedure, color histograms as feature extractor and SVM for the classification task, a database of 1930 images was collected, we used 13 categories of defects described in the SCAA standard of evaluation. Results of classification achieved a 98.8% of overall detection accuracy, therefore the proposed system proved to be effective in classifying physical defects of green coffee beans. Finally a set of conclusions and future works are presented.

This paper describes the development of a system for evaluating the quality of coffee focused on the pre-processing of digital images using an algorithm based on the retinex theory called multi-scale retinex with color restoration (MSRCR). A dataset of images of coffee beans are collected and others techniques for image enhancement are compared, then a color gray-level coocurrence matrix (CGLCM) technique is used for features extraction and a Support Vector Machine (SVM) is used to evaluate results with a set of prepared data, these results shows a good visual quality and better accuracy in classification for MSRCR techniques compared with others, finally conclusions and future works are presented.