Product Quantization for Nearest Neighbor Search Jégou, H; Douze, M; Schmid, C
IEEE transactions on pattern analysis and machine intelligence,
2011-Jan., 2011, 2011-Jan, 2011-01-00, 20110101, 2011-01, Letnik:
33, Številka:
1
Journal Article
Recenzirano
Odprti dostop
This paper introduces a product quantization-based approach for approximate nearest neighbor search. The idea is to decompose the space into a Cartesian product of low-dimensional subspaces and to ...quantize each subspace separately. A vector is represented by a short code composed of its subspace quantization indices. The euclidean distance between two vectors can be efficiently estimated from their codes. An asymmetric version increases precision, as it computes the approximate distance between a vector and a code. Experimental results show that our approach searches for nearest neighbors efficiently, in particular in combination with an inverted file system. Results for SIFT and GIST image descriptors show excellent search accuracy, outperforming three state-of-the-art approaches. The scalability of our approach is validated on a data set of two billion vectors.
Aggregating Local Image Descriptors into Compact Codes Jegou, H.; Perronnin, F.; Douze, M. ...
IEEE transactions on pattern analysis and machine intelligence,
09/2012, Letnik:
34, Številka:
9
Journal Article, Conference Proceeding
Recenzirano
Odprti dostop
This paper addresses the problem of large-scale image search. Three constraints have to be taken into account: search accuracy, efficiency, and memory usage. We first present and evaluate different ...ways of aggregating local image descriptors into a vector and show that the Fisher kernel achieves better performance than the reference bag-of-visual words approach for any given vector dimension. We then jointly optimize dimensionality reduction and indexing in order to obtain a precise vector comparison as well as a compact representation. The evaluation shows that the image representation can be reduced to a few dozen bytes while preserving high accuracy. Searching a 100 million image data set takes about 250 ms on one processor core.
We address the problem of image search on a very large scale, where three constraints have to be considered jointly: the accuracy of the search, its efficiency, and the memory usage of the ...representation. We first propose a simple yet efficient way of aggregating local image descriptors into a vector of limited dimension, which can be viewed as a simplification of the Fisher kernel representation. We then show how to jointly optimize the dimension reduction and the indexing algorithm, so that it best preserves the quality of vector comparison. The evaluation shows that our approach significantly outperforms the state of the art: the search accuracy is comparable to the bag-of-features approach for an image representation that fits in 20 bytes. Searching a 10 million image dataset takes about 50ms.
Attributes were recently shown to give excellent results for category recognition. In this paper, we demonstrate their performance in the context of image retrieval. First, we show that retrieving ...images of particular objects based on attribute vectors gives results comparable to the state of the art. Second, we demonstrate that combining attribute and Fisher vectors improves performance for retrieval of particular objects as well as categories. Third, we implement an efficient coding technique for compressing the combined descriptor to very small codes. Experimental results on the Holidays dataset show that our approach significantly outperforms the state of the art, even for a very compact representation of 16 bytes per image. Retrieving category images is evaluated on the "web-queries" dataset. We show that attribute features combined with Fisher vectors improve the performance and that combined image features can supplement text features.
With the advent of larger image classification datasets such as ImageNet, designing scalable and efficient multi-class classification algorithms is now an important challenge. We introduce a new ...scalable learning algorithm for large-scale multi-class image classification, based on the multinomial logistic loss and the trace-norm regularization penalty. Reframing the challenging non-smooth optimization problem into a surrogate infinite-dimensional optimization problem with a regular ℓ 1 -regularization penalty, we propose a simple and provably efficient accelerated coordinate descent algorithm. Furthermore, we show how to perform efficient matrix computations in the compressed domain for quantized dense visual features, scaling up to 100,000s examples, 1,000s-dimensional features, and 100s of categories. Promising experimental results on the "Fungus", "Ungulate", and "Vehicles" subsets of ImageNet are presented, where we show that our approach performs significantly better than state-of-the-art approaches for Fisher vectors with 16 Gaussians.
Transtracheal oxygen therapy (TTOT) improves the efficiency of oxygen delivery and overcomes the discomfort associated with nasal oxygen delivery in patients on long-term oxygen therapy (LTOT). In ...addition, TTOT improves compliance and quality of life, and may reduce morbidity. Experience with TTOT in Europe is, however, scarce and the safety of TTOT has not yet been completely determined. These were reasons for testing the acceptance, efficacy and safety of TTOT. Patients were selected on the basis of the accepted indications and contraindications for TTOT. In 75 patients (48 males and 27 females) the mean follow-up time was 16 (range 0.5-51.5) months. Compared to nasal cannulae, TTOT caused a reduction in the oxygen flow rate of 47 (33-60)% at rest and a significant increase in the number of hours that oxygen was used. All patients on TTOT used oxygen for at least 20 h x day(-1). Most patients saw the procedure as a minor intervention. It was usually performed on an out-patient basis. In 34 patients, 51 mainly minor complications were seen, and most of these occurred in the first 10 patients. After precautions had been taken, complications occurred less frequently. No patient needed to be hospitalized because of a complication. TTOT had to be stopped in two patients; nevertheless, all patients preferred TTOT to the nasal cannulae. We conclude that transtracheal oxygen therapy is an effective and safe alternative to nasal oxygen administration, provided that it is restricted to a well-defined group of patients and applied by a motivated and experienced group of physicians.
Objectif Devant les difficultés des infirmières libérales pour coordonner chez les patients diabétiques insulino-traités leurs soins avec les médecins traitants, les diabétologues et l’USAD (Unité de ...Soins Ambulatoire de Diabétologie) de… ; j’ai organisé quatre réunions de réflexion avec une douzaine d’infirmières libérales. Matériels et méthodes Il en est ressorti l’intérêt de créer un carnet de liaison pour les infirmières qui prennent en charge l’insulinothérapie à domicile des patients non autonomes. Ce carnet permet d’échanger des informations complètes sur les traitements oraux et injectables ainsi que sur les protocoles thérapeutiques des différents schémas d’insuline, les objectifs glycémiques, les prescriptions médicales pour un rajout ponctuel d’insuline. Il précise les consultations et prélèvements prévus ou à prévoir. Il contient également plusieurs pages destinées à poser et répondre à des questions diverses et variées, à noter des transmissions et des conseils personnalisés. Après sa conception en partenariat avec les infirmières libérales, ce carnet a été présenté à toute l’équipe médicale et paramédicale de l’USAD qui l’a approuvé. Résultats Il va être testé par douze infirmières libérales (dix carnets chacune) pour valider qu’il est possible d’améliorer, grâce à son usage, la prise en charge du patient diabétique en coordonnant les soins et en favorisant leurs échanges avec les médecins traitants, les diabétologues et les infirmières de l’USAD. Conclusion Le carnet tel qu’il a été conçu sera affiché avec le poster.
We propose a 2D approach for augmented reality (AR) applications where the real scene is modelled as a static panorama. We adapted a sparse tracking method based on homographies to track the ...orientation and zooming parameters of the camera during a video sequence. AR scenarii (synthetic object insertion, real object or character extraction) can be performed in arbitrary static environments (from wide outdoor scenes to virtually augmented desktops or conference rooms).
Accuracy of oxygen delivery by liquid oxygen canisters Kampelmacher, M J; Cornelisse, P B; Alsbach, G P ...
European respiratory journal/The European respiratory journal,
07/1998, Letnik:
12, Številka:
1
Journal Article
Recenzirano
Odprti dostop
The oxygen flow rate delivered by liquid oxygen canisters may be less than intended, owing to inaccuracies of the set flow rates and/or as a result of the outflow resistance caused by the humidifier, ...oxygen tubing, delivery or conserving device. The aim of this study was to investigate the accuracy of oxygen delivery by liquid oxygen canisters at different flow rates and levels of outflow resistance. Four stationary and 18 portable liquid oxygen canisters from three manufacturers were tested. All flows were measured using a Timemeter RT 200 Calibration Analyser. An adjustable obstruction was used to calculate the effect of the outflow resistance on the delivered flow rate. The measured and set flow rates of both stationary and portable canisters were strongly correlated. Expressed as a percentage of the set flow rate, the measured flow rate of the canisters varied from 36-128%, with the lowest values at flow rates <1 L x min(-1). Sixty-two (26%) of the measured flow rates differed > or =10% from the set flow rate. A difference of 0.5-1.0 L x min(-1) occurred in 16 (7%) of the measurements, but only at set flow rates >2 L x min(-1). Irrespective of the set flow rate, the measured flow rate was hardly affected by the outflow resistance. We conclude that the accuracy of oxygen delivery by liquid oxygen canisters depends on the inaccuracy of the set flow rates rather than on the outflow resistance, even at high flow rates. Although the differences between the measured and set flow rates were mostly small, they may have clinical significance, particularly in patients with chronic obstructive pulmonary disease. To prevent inaccuracy, monthly checks of the canisters by the oxygen provider at the patient's home are more important than yearly maintenance.
In patients prescribed long-term oxygen therapy (LTOT), compliance is often poor. Both patient- and treatment-related factors seem to be involved. As a base for improvements in LTOT, the ...characteristics and complaints of LTOT patients were investigated.
A survey was set up in a random sample of clients of the largest oxygen company in the Netherlands. Patients were selected if they were ≥18 years old, had a phone and if they had had oxygen equipment for ≥6 months. All patients were visited at home by a medical student. Data are presented for a total of 528 patients (response rate 62%). The typical LTOT patient was a 70-year-old male with chronic obstructive pulmonary disease (COPD), who had had oxygen equipment for 3·5 years and who used oxygen cylinders and nasal cannulae for 13 h day
−1. Twenty percent of the patients still smoked. Although LTOT was prescribed in 80% of the patients by a chest physician, prescription was often inadequate. Only 33% of the patients were informed adequately about the therapy. Twenty percent of the patients used oxygen for fewer hours per day than prescribed. Non-compliant patients were mainly men (
P=0·006) and more often ashamed of their therapy (
P=0·023) than compliant patients. The blood oxygen level was monitored regularly in 73% of the patients. Most complaints concerned the oxygen equipment, especially the concentrator. The single most important complaint had to do with restricted autonomy. Only 19% of the patients had no complaints at all.
It is concluded that LTOT should be improved with regard to the education, motivation and monitoring of patients. The prescribing physician needs to be included in an education programme. Given the numerous problems these patients experience, LTOT should be improved in particular with regard to equipment convenience.