Markov Random Fields with Asymmetric Interactions for Modelling Spatial Context in Structured Scene Labelling

In this paper we propose a Markov random field with asymmetric Markov parameters to model the spatial and topological relationships between objects in structured scenes. The field is formulated in terms of conditional probabilities learnt from a set of training images. A locally consistent labelling of new scenes is achieved by relaxing the Markov random field directly using these conditional probabilities. We evaluate our model on a varied collection of several hundred hand-segmented images of buildings. The incorporation of spatial information is shown to improve greatly the performance of some trivial classifiers.     

Intron position affects expression from the tpi promoter in rice

A series of promoter-GUS fusion constructs containing a portion of the rice triosephosphate isomerase (tpi) promoter, the first tpi intron, and the gene encoding bacterial beta-glucuronidase (GUS) were made. These constructs were electroporated into rice protoplasts and transient expression was monitored. Inclusion of the first intron from the rice tpi gene enhanced expression of the GUS gene from the tpi promoter when it was placed 5' of the GUS gene. When the tpi intron was placed in the 3'-untranslated region no enhancement of GUS gene expression was observed, indicating the importance of position in intron-mediated enhancement of gene expression.     

Pallidal lesions. Structural and functional magnetic resonance imaging

OBJECTIVE: To study noninvasively the functional anatomy and pathophysiologic characteristics of the globus pallidus external (GPe) and internal (GPi) divisions. DESIGN: Structural and functional neuroimaging using high-resolution magnetic resonance imaging. SETTING: University medical center research facility. SUBJECTS. Seven patients with pallidal lesions, 4 with an akinetic-rigid syndrome and 3 with a dystonic syndrome, and 15 age-matched volunteers. MAIN OUTCOME MEASURES: T2-weighted anatomical magnetic resonance imaging and number of activated voxels in the GP during rapid supination and pronation of the hand. RESULTS: T2-weighted images showed hyperintense bilateral lesions in the GP of all patients. Patients with dystonic syndromes had isolated lesions in the GPi. Patients with signs of akinetic-rigid syndromes showed abnormalities in the GPe or in central portions of the GP (GPc). Patients with lesions in both parts of the GP had akinetic-rigid or dystonic syndromes. All patients showed activation in the areas of the lesions. The number of activated voxels in the GP was significantly smaller (P < .005, Wilcoxon signed rank test) in patients than in control subjects. Activation of the GP was predominantly contralateral to the moving hand. CONCLUSIONS: Lesions in the GPi result in a loss of inhibitory pallidal projections to the thalamus, which may explain the hyperkinetic signs. Lesions in the GPe lead to an increased inhibition of the thalamus, which may explain the hypokinetic signs. Neuronal activation in lesion sites suggests the presence of remaining functionally vital tissue.