Yahoo!

About: Yahoo! is a company organization based out in London, United Kingdom. It is known for research contribution in the topics: Population & Web search query. The organization has 26749 authors who have published 29915 publications receiving 732583 citations. The organization is also known as: Yahoo! Inc. & Maudwen-Yahoo! Inc.

PET/CT in Lymphoma: Prospective Study of Enhanced Full-Dose PET/CT Versus Unenhanced Low-Dose PET/CT

Abstract: PET/CT combines functional and morphologic data and increases diagnostic accuracy in a variety of malignancies. This study prospectively compares the agreement between contrast-enhanced full-dose PET/CT and unenhanced low-dose PET/CT in lesion detection and initial staging of Hodgkin9s disease and non-Hodgkin9s lymphoma. Methods: Forty-seven biopsy-proven lymphoma patients underwent a 18F-FDG PET/CT study that included unenhanced low-dose CT and enhanced full-dose CT for initial staging. Patients who had undergone previous diagnostic CT for initial staging were excluded. For every patient, each modality of PET/CT images was evaluated by either of 2 pairs of readers, with each pair comprising 1 experienced radiologist and 1 experienced nuclear physician. While evaluating one of the 2 types of PET/CT, the readers were unaware of the results of the other type. Lesion detection, number of sites affected in each anatomic region, and disease stage were assessed. Agreement between techniques was determined by the κ-statistic, and discordances were studied by the McNemar test. Clinical, analytic, histopathologic, diagnostic CT, and PET data; data from other imaging techniques; and follow-up data constituted the reference standard. Results: For region-based analysis, no significant differences were found between unenhanced low-dose PET/CT and contrast-enhanced full-dose PET/CT, although full-dose PET/CT showed fewer indeterminate findings and a higher number of extranodal sites affected than did low-dose PET/CT. Agreement between the 2 types of PET/CT was almost perfect for disease stage (κ = 0.92; P

Media object metadata engine configured to determine relationships between persons and brands

Abstract: A media object, such as an image file, a video file, or an audio file, is analyzed to determine relationships between brands having representations captured in the media object, and persons associated with the media object (e.g., persons captured in the media object and/or a person that captured the media object). A representation of a brand captured in a media object is detected. The media object is analyzed to determine at least one indicator of a relation between the brand and a person associated with the media object. A relationship between the brand and person is predicted based at least on the determined at least one relation indicator. The media object may be monetized in various ways, such as by directing advertisements (e.g., advertisements related to the detected brand) to persons associated with the media object, and/or to persons having social connections to the persons associated with the media object.

Automatically matching advertisements to media files

Abstract: The present invention relates to a system and method for selecting advertisement for delivery over a network in response to requests received from remote computing devices. In one respect, the present invention includes a method and system for automatically matching an advertisement with a media file, such as a podcast episode, when the media file has been requested by a consumer. Aspects of the present invention allow automatic selection of advertisements after the creation of the media file, potentially without any interaction between the creator and the advertiser.

Fast locality-sensitive hashing

Abstract: Locality-sensitive hashing (LSH) is a basic primitive in several large-scale data processing applications, including nearest-neighbor search, de-duplication, clustering, etc. In this paper we propose a new and simple method to speed up the widely-used Euclidean realization of LSH. At the heart of our method is a fast way to estimate the Euclidean distance between two d-dimensional vectors; this is achieved by the use of randomized Hadamard transforms in a non-linear setting. This decreases the running time of a (k, L)-parameterized LSH from O(dkL) to O(dlog d + kL). Our experiments show that using the new LSH in nearest-neighbor applications can improve their running times by significant amounts. To the best of our knowledge, this is the first running time improvement to LSH that is both provable and practical.

Scalable triangulation-based logo recognition

Abstract: We propose a scalable logo recognition approach that extends the common bag-of-words model and incorporates local geometry in the indexing process. Given a query image and a large logo database, the goal is to recognize the logo contained in the query, if any. We locally group features in triples using multi-scale Delaunay triangulation and represent triangles by signatures capturing both visual appearance and local geometry. Each class is represented by the union of such signatures over all instances in the class. We see large scale recognition as a sub-linear search problem where signatures of the query image are looked up in an inverted index structure of the class models. We evaluate our approach on a large-scale logo recognition dataset with more than four thousand classes.