Evaluation of transfer learning of pre-trained CNNs applied to breast cancer detection on infrared images.

Breast cancer accounts for the highest number of female deaths worldwide. Early detection of the disease is essential to increase the chances of treatment and cure of patients. Infrared thermography has emerged as a promising technique for diagnosis of the disease due to its low cost and that it does not emit harmful radiation, and it gives good results when applied in young women. This work uses convolutional neural networks in a database of 440 infrared images of 88 patients, classifying them into two classes: normal and pathology. During the training of the networks, we use transfer learning of the following convolutional neural network architectures: AlexNet, GoogLeNet, ResNet-18, VGG-16, and VGG-19. Our results show the great potential of using deep learning techniques combined with infrared images in the aid of breast cancer diagnosis.

A study of publication trajectories of the Brazilian Computer Science community.

The average faculty productivity have been described as a rapid rise-short peak-gradual decline pattern. Way et al. (2017) have studied this pattern for faculty careers in Computer Science in North America using a piecewise linear model. In this paper, we use a similar methodology and study trajectories (N = 20655) of the Brazilian Computer Science community. First, we have evaluated how the median publication count of researchers is related to institution prestige and public vs. private administration. Second, we have studied how the annual publication rates have increased and its variation according to prestige ranks of institutions. Third, we have found the average trajectory can indeed be described as the canonical rapid increase and slower decrease in productivity. For individual trajectories of senior researchers we have observed only 4.5% of trajectories are well explained by the conventional narrative of rapid rise and gradual decline model. We also have found polynomial models of degrees 1 to 3 explain almost 63.1% of trajectories. The rest of trajectories are considered unstable and not well explained by neither of approaches.