Abstract: Endoscopic estimation of the degree of stenosis in central airway obstruction is subjective and highly variable. Objective: To determine the benefits of using SENSA (System for Endoscopic Stenosis Assessment), an image-based computational software, for obtaining objective stenosis index (SI) measurements among a group of expert bronchoscopists and general pulmonologists. Methods: A total of 7 expert bronchoscopists and 7 general pulmonologists were enrolled to validate SENSA usage. The SI obtained by the physicians and by SENSA were compared with a reference SI to set their precision in SI computation. We used SENSA to efficiently obtain this reference SI in 11 selected cases of benign stenosis. A Web platform with three user-friendly microtasks was designed to gather the data. The users had to visually estimate the SI from videos with and without contours of the normal and the obstructed area provided by SENSA. The users were able to modify the SENSA contours to define the reference SI using morphometric bronchoscopy. Results: Visual SI estimation accuracy was associated with neither bronchoscopic experience (p = 0.71) nor the contours of the normal and the obstructed area provided by the system (p = 0.13). The precision of the SI by SENSA was 97.7% (95% CI: 92.4-103.7), which is significantly better than the precision of the SI by visual estimation (p < 0.001), with an improvement by at least 15%. Conclusion: SENSA provides objective SI measurements with a precision of up to 99.5%, which can be calculated from any bronchoscope using an affordable scalable interface. Providing normal and obstructed contours on bronchoscopic videos does not improve physicians' visual estimation of the SI.

Abstract: Word-spotting methods based on the Bag-ofVisual-Words framework have demonstrated a good retrieval performance even when used in a completely unsupervised manner. Although unsupervised approaches are suitable for
large document collections due to the cost of acquiring labeled data, these methods also present some drawbacks. For instance, having to train a suitable “codebook” for a certain dataset has a high computational cost. Therefore, in
this paper we present a database agnostic codebook which is trained from synthetic data. The aim of the proposed approach is to generate a codebook where the only information required is the type of script used in the document. The use of synthetic data also allows to easily incorporate semantic
information in the codebook generation. So, the proposed method is able to determine which set of codewords have a semantic representation of the descriptor feature space. Experimental results show that the resulting codebook attains a state-of-the-art performance while having a more compact representation.

Abstract: In this work, we explore the use of images from different spectral bands to classify defects in melamine faced panels, which could appear through the production process. Through experimental evaluation, we evaluate the use of images from the visible (VS), near-infrared (NIR), and long wavelength infrared (LWIR), to classify the defects using a feature descriptor learning approach together with a support vector machine classifier. Two descriptors were evaluated, Extended Local Binary Patterns (E-LBP) and SURF using a Bag of Words (BoW) representation. The evaluation was carried on with an image set obtained during this work, which contained five different defect categories that currently occurs in the industry. Results show that using images from beyond the visual spectrum helps to improve classification performance in contrast with a single visible spectrum solution.

Abstract: Previous works on sequential learning address the problem of forgetting in discriminative models. In this paper we consider the case of generative models. In particular, we investigate generative adversarial networks (GANs) in the task of learning new categories in a sequential fashion. We first show that sequential fine tuning renders the network unable to properly generate images from previous categories (ie forgetting). Addressing this problem, we propose Memory Replay GANs (MeRGANs), a conditional GAN framework that integrates a memory replay generator. We study two methods to prevent forgetting by leveraging these replays, namely joint training with replay and replay alignment. Qualitative and quantitative experimental results in MNIST, SVHN and LSUN datasets show that our memory replay approach can generate competitive images while significantly mitigating the forgetting of previous categories.

Abstract: In this dissertation, we explore different ways to perform human action recognition in video clips. We focus on data efficiency, proposing new approaches that alleviate the need for laborious and time-consuming manual data annotation. In the first part of this dissertation, we start by analyzing previous state-of-the-art models, comparing their differences and similarities in order to pinpoint where their real strengths come from. Leveraging this information, we then proceed to boost the classification accuracy of shallow models to levels that rival deep neural networks. We introduce hybrid video classification architectures based on carefully designed unsupervised representations of handcrafted spatiotemporal features classified by supervised deep networks. We show in our experiments that our hybrid model combine the best of both worlds: it is data efficient (trained on 150 to 10,000 short clips) and yet improved significantly on the state of the art, including deep models trained on millions of manually labeled images and videos. In the second part of this research, we investigate the generation of synthetic training data for action recognition, as it has recently shown promising results for a variety of other computer vision tasks. We propose an interpretable parametric generative model of human action videos that relies on procedural generation and other computer graphics techniques of modern game engines. We generate a diverse, realistic, and physically plausible dataset of human action videos, called PHAV for “Procedural Human Action Videos”. It contains a total of 39,982 videos, with more than 1,000 examples for each action of 35 categories. Our approach is not limited to existing motion capture sequences, and we procedurally define 14 synthetic actions. We then introduce deep multi-task representation learning architectures to mix synthetic and real videos, even if the action categories differ. Our experiments on the UCF-101 and HMDB-51 benchmarks suggest that combining our large set of synthetic videos with small real-world datasets can boost recognition performance, outperforming fine-tuning state-of-the-art unsupervised generative models of videos.

Abstract: Few-shot learning has become essential for producing models that generalize from few examples. In this work, we identify that metric scaling and metric task conditioning are important to improve the performance of few-shot algorithms. Our analysis reveals that simple metric scaling completely changes the nature of few-shot algorithm parameter updates. Metric scaling provides improvements up to 14% in accuracy for certain metrics on the mini-Imagenet 5-way 5-shot classification task. We further propose a simple and effective way of conditioning a learner on the task sample set, resulting in learning a task-dependent metric space. Moreover, we propose and empirically test a practical end-to-end optimization procedure based on auxiliary task co-training to learn a task-dependent metric space. The resulting few-shot learning model based on the task-dependent scaled metric achieves state of the art on mini-Imagenet. We confirm these results on another few-shot dataset that we introduce in this paper based on CIFAR100.

Abstract: One of the fundamental properties of a salient object region is its contrast
with the immediate context. The problem is that numerous object regions
exist which potentially can all be salient. One way to prevent an exhaustive
search over all object regions is by using object proposal algorithms. These
return a limited set of regions which are most likely to contain an object. Several saliency estimation methods have used object proposals. However, they focus on the saliency of the proposal only, and the importance of its immediate context has not been evaluated.
In this paper, we aim to improve salient object detection. Therefore, we extend object proposal methods with context proposals, which allow to incorporate the immediate context in the saliency computation. We propose several saliency features which are computed from the context proposals. In the experiments, we evaluate five object proposal methods for the task of saliency segmentation, and find that Multiscale Combinatorial Grouping outperforms the others. Furthermore, experiments show that the proposed context features improve performance, and that our method matches results on the FT datasets and obtains competitive results on three other datasets (PASCAL-S, MSRA-B and ECSSD).

Abstract: Optical Music Recognition refers to the task of transcribing the image of a music score into a machine-readable format. Many music scores are written in a single staff, and therefore, they could be treated as a sequence. Therefore, this work explores the use of Long Short-Term Memory (LSTM) Recurrent Neural Networks for reading the music score sequentially, where the LSTM helps in keeping the context. For training, we have used a synthetic dataset of more than 40000 images, labeled at primitive level. The experimental results are promising, showing the benefits of our approach.