<abstract>
<p>Currently, the gastric cancer is the source of the high mortality rate where it is diagnoses from the stomach and esophagus tests. To this end, the whole of studies in the analysis of cancer are built on AI (artificial intelligence) to develop the analysis accuracy and decrease the danger of death. Mostly, deep learning methods in images processing has made remarkable advancement. In this paper, we present a method for detection, recognition and segmentation of gastric cancer in endoscopic images. To this end, we propose a deep learning method named GAS-Net to detect and recognize gastric cancer from endoscopic images. Our method comprises at the beginning a preprocessing step for images to make all images in the same standard. After that, the GAS-Net method is based an entire architecture to form the network. A union between two loss functions is applied in order to adjust the pixel distribution of normal/abnormal areas. GAS-Net achieved excellent results in recognizing lesions on two datasets annotated by a team of expert from several disciplines (Dataset1, is a dataset of stomach cancer images of anonymous patients that was approved from a private medical-hospital clinic, Dataset2, is a publicly available and open dataset named HyperKvasir <xref ref-type="bibr" rid="b1">[1]</xref>). The final results were hopeful and proved the efficiency of the proposal. Moreover, the accuracy of classification in the test phase was 94.06%. This proposal offers a specific mode to detect, recognize and classify gastric tumors.</p>
</abstract>
<p><strong>Abstract.</strong> In this work, we propose an efficient way of web video annotation in soccer domain. To achieve this, it is necessary to enjoy different architectures of deep learning. We aim at realising a system of annotation able to recognise several events from information of the object that is the ball in our case, in order to fuse them as a part of actions in video. We propose to use Deep Neural Network (DNN) to detect ball and actions. However, Mask R-CNN can play a very important role for features extracted as an output using a training network on ImageNet dataset. The Mask R-CNN is chosen as a method using different CNN as backbone (convolutional Neural Network) ResNet50, ResNet101 and ResNet152, VGG16, VGG 19. We experimentally verify the effectiveness of the proposed method in the test phase.</p>
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.