Relevance feedback for CBIR: a new approach based on probabilistic feature weighting with positive and negative examples.

In content-based image retrieval, understanding the user's needs is a challenging task that requires integrating him in the process of retrieval. Relevance feedback (RF) has proven to be an effective tool for taking the user's judgement into account. In this paper, we present a new RF framework based on a feature selection algorithm that nicely combines the advantages of a probabilistic formulation with those of using both the positive example (PE) and the negative example (NE). Through interaction with the user, our algorithm learns the importance he assigns to image features, and then applies the results obtained to define similarity measures that correspond better to his judgement. The use of the NE allows images undesired by the user to be discarded, thereby improving retrieval accuracy. As for the probabilistic formulation of the problem, it presents a multitude of advantages and opens the door to more modeling possibilities that achieve a good feature selection. It makes it possible to cluster the query data into classes, choose the probability law that best models each class, model missing data, and support queries with multiple PE and/or NE classes. The basic principle of our algorithm is to assign more importance to features with a high likelihood and those which distinguish well between PE classes and NE classes. The proposed algorithm was validated separately and in image retrieval context, and the experiments show that it performs a good feature selection and contributes to improving retrieval effectiveness.     

A new approach to modelling of single droplet drying

A model was developed to predict the change in droplet mass and temperature when it is exposed to hot air, as in spray drying of droplets containing solids. The droplet was assumed first to undergo sensible rapid heating with no mass change. Then the droplet experiences some shrinkage, with no temperature change but rapid mass losses, followed by a period of crust formation with a significant change in droplet mass and temperature and finally a short period of sensible heating of the dried particle. The model, unlike previous models, accounts for shrinkage and for the temperature distribution in the droplet. It provided a good prediction for the change in droplet temperature and mass for some of the experimental measurements available in the literatures.     

Gastric ECL-cell hyperplasia produces enhanced basal and stimulated gastric acid output but not gastric erosion formation in the rat

Cervical involvement is one of the major prognostic factors in carcinoma of the endometrium confined to the uterus. The purpose of this study was to determine whether intrauterine ultrasound with a high-frequency miniature probe can depict the degree of cervical involvement of the disease. Thirty-two women with endometrial carcinoma underwent preoperative transvaginal and intrauterine sonography. By both scans, the degree of cervical involvement was prospectively evaluated. Sonograms were compared with the findings from histologic examination. Intrauterine sonography was completed in 30 of the 32 patients. In these 30 patients, the degree of cervical involvement (none, endocervical gland, or cervical stroma) based on transvaginal scan was correct in 23 cases (77%), and that based on intrauterine scan was correct in 26 cases (87%). Three tumors with endocervical glandular involvement were correctly diagnosed by intrauterine sonography, whereas they were incorrectly diagnosed by transvaginal scan. The specificity and positive predictive value of intrauterine sonography for the assessment of the presence of cervical stromal invasion are 100% (26/26 and 3/3, respectively). Although this study is preliminary, our experience with intrauterine sonography shows that it has potential for assessing cervical stromal invasion in endometrial carcinoma.