Abstract

Abstract
This study explored the viability of out-the-box, pre-trained ConvNet models as a tool to generate features for large-scale classification tasks. A juxtaposition with generative methods for vocabulary generation was drawn. Both methods were chosen in an attempt to integrate other datasets (transfer learning) and unlabelled data, respectively. Both methods were used together, studying the viability of a ConvNet model to estimate category labels of unlabelled images. All experiments pertaining to this study were carried out over a two-class set, later expanded into a 5-category dataset. The pre-trained models used were obtained from the Caffe Model Zoo. The study showed that the pre-trained model achieved best results for the binary dataset, with an accuracy of 0.945. However, for the 5-class dataset, generative vocabularies outperformed the ConvNet (0.91 vs. 0.861). Furthermore, when replacing labelled images with unlabelled ones during training, acceptable accuracy scores were obtained (as high as 0.903). Additionally, it was observed that linear kernels perform particularly well when utilized with generative models. This was especially relevant when compared to ConvNets, which require days of training even when utilizing multiple GPUs for computations.

Abstract
In this paper we present the INESC Key Detection (IKD) system which incorporates a novel method for dynamically biasing key mode estimation using the spatial displacement of beat-synchronous Tonal Interval Vectors (TIVs). We evaluate the performance of the IKD system at finding the global key on three annotated audio datasets and using three key-defining profiles. Results demonstrate the effectiveness of the mode bias in favoring either the major or minor mode, thus allowing users to fine tune this variable to improve correct key estimates on style-specific music datasets or to balance predictions across key modes on unknown input sources.

Abstract
We present a hierarchical harmonic mixing method for assisting users in the process of music mashup creation. Our main contributions are metrics for computing the harmonic compatibility between musical audio tracks at small- and large-scale structural levels, which combine and reassess existing perceptual relatedness (i.e., chroma vector similarity and key affinity) and dissonance-based approaches. Underpinning our harmonic compatibility metrics are harmonic indicators from the perceptually-motivated Tonal Interval Space, which we adapt to describe musical audio. An interactive visualization shows hierarchical harmonic compatibility viewpoints across all tracks in a large musical audio collection. An evaluation of our harmonic mixing method shows our adaption of the Tonal Interval Space robustly describes harmonic attributes of musical instrument sounds irrespective of timbral differences and demonstrates that the harmonic compatibility metrics comply with the principles embodied in Western tonal harmony to a greater extent than previous approaches. © 2018, Springer Nature Switzerland AG.

Abstract
Automatic segmentation of breast is an important step in the context of providing a planning tool for breast cancer conservative treatment, being important to segment completely the breast region in an objective way; however, current methodologies need user interaction or detect breast contour partially. In this paper, we propose a methodology to detect the complete breast contour, including the pectoral muscle, using multi-modality data. Exterior contour is obtained from 3D reconstructed data acquired from low-cost RGB-D sensors, and the interior contour (pectoral muscle) is obtained from Magnetic Resonance Imaging (MRI) data. Quantitative evaluation indicates that the proposed methodology performs an acceptable detection of breast contour, which is also confirmed by visual evaluation.

Abstract
The segmentation of the anterior fascia of the rectus abdominis muscle is an important step towards the analysis of abdominal vasculature. It may advance Computer Aided Detection tools that support the activity of clinicians who study vessels for breast reconstruction using the Deep Inferior Epigastric Perforator flap. In this paper, we propose a two-fold methodology to detect the anterior fascia in Computerized Tomographic Angiography volumes. First, a slice-wise thresholding is applied and followed by a post-processing phase. Finally, an interpolation framework is used to obtain a final smooth fascia detection. We evaluated our method in 20 different volumes, by calculating the mean Euclidean distance to manual annotations, achieving subvoxel error.