Learn More
Users often have very specific visual content in mind that they are searching for. The most natural way to communicate this content to an image search engine is to use key-words that specify various properties or attributes of the content. A naive way of dealing with such multi-attribute queries is the following: train a classifier for each attribute(More)
Object detection is a challenging task in visual understanding domain, and even more so if the supervision is to be weak. Recently, few efforts to handle the task without expensive human annotations is established by promising deep neural network. A new architecture of cascaded networks is proposed to learn a convolutional neural network (CNN) under such(More)
In this paper we evaluate the quality of the activation layers of a convolutional neural network (CNN) for the generation of object proposals. We generate hypotheses in a sliding-window fashion over different activation layers and show that the final convolutional layers can find the object of interest with high recall but poor localization due to the(More)
A total of 52 jaundiced elderly patients who had malignant obstruction of the distal common bile duct and who required palliative biliary decompression were randomized to receive either an endoscopically placed biliary endoprosthesis (10 French gauge) or conventional surgical bypass. Patients within the two treatment groups were well matched and 51 were(More)
The use of elective tracheostomy in major head and neck surgery is well established, although practice varies between units. There is no published method that reliably predicts the need for tracheostomy. This paper describes the development of a surgical scoring system designed to achieve that aim. The system was devised using data obtained retrospectively(More)
In this paper, a new method for generating object and action proposals in images and videos is proposed. It builds on activations of different convolutional layers of a pretrained CNN, combining the localization accuracy of the early layers with the high informativeness (and hence recall) of the later layers. To this end, we build an inverse cascade that,(More)
A patient who had a 50% gastrectomy with a Billroth II gastrojejunostomy one and a half years previously, complained of recent severe weakness as the only symptom was found to have an iron deficiency anemia with a periumbilical mass. A gastrointestinal series showed a soft tissue density in the epigastric area which, by ultrasonography, was found to be(More)
Tracheal airway pressures were measured via a transduced fibrescope during transtracheal jet ventilation in 10 patients. Ravussin transtracheal jet ventilation catheters were inserted under local anaesthesia. Following induction of general anaesthesia, tracheal airway pressures were measured at three anatomical levels during fibreoptic intubation. Overall(More)
The video and action classification have extremely evolved by deep neural networks specially with two stream CNN using RGB and optical flow as inputs and they present outstanding performance in terms of video analysis. One of the shortcoming of these methods is handling motion information extraction which is done out side of the CNNs and relatively time(More)