Accurate diagnosis of breast cancer in histopathology images is challenging due to the heterogeneity of cancer cell growth as well as of a variety of benign breast tissue proliferative lesions. In this paper, we propose a practical and self-interpretable invasive cancer diagnosis solution. With minimum annotation information, the proposed method mines contrast patterns between normal and malignant images in weaksupervised manner and generates a probability map of abnormalities to verify its reasoning. Particularly, a fully convolutional autoencoder is used to learn the dominant structural patterns among normal image patches. Patches that do not share the characteristics of this normal population are detected and analyzed by one-class support vector machine and 1-layer neural network.We apply the proposed method to a public breast cancer image set. Our results, in consultation with a senior pathologist, demonstrate that the proposed method outperforms existing methods. The obtained probability map could benefit the pathology practice by providing visualized verification data and potentially leads to a better understanding of data-driven diagnosis solutions.

The ever decreasing area of forests has lead to environmental and economical challenges and has brought with it a renewed interest in developing methodologies that quantify the extent of deforestation and reforestation. In this study we analyzed the deforested areas of the Apuseni Mountains, which has been under economic pressure in recent years and resulted in widespread deforestation as a means of income. Deforested surface dynamics modeling was based on images contained in the Global Forest Database, provided by the Department of Geographical Sciences at Maryland University between 2000 and 2014. The results of the image particle analysis and modelling were based on Total Area (ha), Count of patches and Average Size whereas deforested area distribution was based on the Local Connected Fractal Dimension, Fractal Fragmentation Index and Tug-of-War Lacunarity as indicators of forest fragmentation or heterogeneity. The major findings of the study indicated a reduction of the tree cover area by 3.8%, an increase in fragmentation of 17.7% and an increase in heterogeneity by 29%, while fractal connectivity decreased only by 0.1%. The fractal and particle analysis showed a clustering of forest loss areas with an average increase from 1.1 to 3.0 ha per loss site per year. In conclusion, the fractal and particle analysis provide a relevant methodological framework to further our understanding of the spatial effects of economic pressure on forestry.

Corticotropin-releasing factor (CRF), urocortin, sauvagine and urotensin I form the CRF family. These peptides bind with different affinities to two subtypes of CRF receptor (CRFR), CRFR1 and CRFR2. The latter exists as two splice variants, the neuronal CRFR2a. and the peripheral CRFR2. CRFR is a G protein-dependent receptor which acts mainly through Gs enhancing cAMP production. However, CRFR1 expressed in neutrophils of the spleen in response to immunologic stimulation and psychological stress does not seem to function through Gs, as indicated by the inability of CRF to stimulate the cAMP production of CRFR1+ neutrophils. Besides the two receptors, a 37 kD CRF binding protein (CRF-BP) binds several CRF peptides with high affinity. CRFR and CRF-BP do not share a common amino acid sequence representing the ligand binding site. In view of the unusually slow offrate of CRF-BP, it is proposed that CRF-BP provides an efficient uptake of free extracellular CRF. Thus, the time of exposure of CRFR to CRF or urocortin can be limited. At this time, the fate of the ligand CRF­ BP complex is unclear. CRFR1 is not only involved in the hypophyseal stimulation of corticotropin release, but hippocampal CRFR1 mediates enhancement of stress-induced learning. CRFR1 may also be involved in basic anxiety. In contrast, at least in the mouse, CRFR2 of the lateral intermediate septum mediates tonic impairment of learning. In response to stressful stimuli or after local injection of high CRF doses, CRFR2 mediates anxiety. Effects requiring CRFR2 can be blocked specifically by the recently developed peptidic antagonist antisauvagine- 30.

A specific polyclonal Ab against the N-terminal domain of corticotropin-releasing factor (CRF) receptor, type 1 (CRF-R1), was employed to an immunohistochemical analysis of the spleen from naive mice and mice exposed to an immune challenge. Cell types stained with anti-CRF-R1 Ab were identified by their nuclear shapes and colocalization with the cell type-specific markers ER-MP58, ER-MP20, Moma-1, Moma 2, anti-CD3e mAbs, and anti-Ig Ab. Only a few clusters of CRF-R1+ cells were found in spleen sections of naive mice at sites typical for granulopoietic islands. However, a 17-fold increase in the mean number of CRF-R1+ cells was noted within hours following a challenge of acute systemic inflammation induced by i.p. administration of LPS. The majority of these cells were identified as mature neutrophils. CRF-R1 was shown to mediate suppression of the IL-1β secretion by these cells. However, at later time points a large number of granulocyte-macrophage precursors was strongly labeled with anti-CRF-R1 Ab. Western blot analysis of splenic membranes from animals treated with LPS revealed a m.w. of approximately 70,000 for CRF-R1. Subcellular staining patterns were suggestive for the predominant localization of CRF-R1 on granule membranes. CRF-R1 mRNA was detected in spleen but not in bone marrow and peripheral blood leukocytes from naive mice. Thus, it was indicated that CRF-R1 was not produced constitutively by mature or immature neutrophils. Its production was rather triggered by inflammatory stimuli.