Abstract: Facial dynamics can be considered as unique signatures for discrimination between people. These have started to become important topic since many devices have the possibility of unlocking using face recognition or verification. In this work, we evaluate the efficacy of the transition frames of video in emotion as compared to the peak emotion frames for identification. For experiments with transition frames we extract features from each frame of the video from a fine-tuned VGG-Face Convolutional Neural Network (CNN) and geometric features from facial landmark points. To model the temporal context of the transition frames we train a Long-Short Term Memory (LSTM) on the geometric and the CNN features. Furthermore, we employ two fusion strategies: first, an early fusion, in which the geometric and the CNN features are stacked and fed to the LSTM. Second, a late fusion, in which the prediction of the LSTMs, trained independently on the two features, are stacked and used with a Support Vector Machine (SVM). Experimental results show that the late fusion strategy gives the best results and the transition frames give better identification results as compared to the peak emotion frames.

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: We address the problem of image translation between domains or modalities for which no direct paired data is available (i.e. zero-pair translation). We propose mix and match networks, based on multiple encoders and decoders aligned in such a way that other encoder-decoder pairs can be composed at test time to perform unseen image translation tasks between domains or modalities for which explicit paired samples were not seen during training. We study the impact of autoencoders, side information and losses in improving the alignment and transferability of trained pairwise translation models to unseen translations. We show our approach is scalable and can perform colorization and style transfer between unseen combinations of domains. We evaluate our system in a challenging cross-modal setting where semantic segmentation is estimated from depth images, without explicit access to any depth-semantic segmentation training pairs. Our model outperforms baselines based on pix2pix and CycleGAN models.

Abstract: Image datasets have heavily been used to build computer vision systems.
These datasets are either manually or automatically labeled, which is a
problem as both labeling methods are prone to errors. To investigate this problem, we use a majority voting ensemble that combines the results from several Convolutional Neural Networks (CNNs). Majority voting ensembles not only enhance the overall performance, but can also be used to estimate the confidence level of each sample. We also examined Softmax as another form to estimate posterior probability. We have designed various experiments with a range of different ensembles built from one or different, or temporal/snapshot CNNs, which have been trained multiple times stochastically. We analyzed CIFAR10, CIFAR100, EMNIST, and SVHN datasets and we found quite a few incorrect
labels, both in the training and testing sets. We also present detailed confidence analysis on these datasets and we found that the ensemble is better than the Softmax when used estimate the per-sample confidence. This work thus proposes an approach that can be used to scrutinize and verify the labeling of computer vision datasets, which can later be applied to weakly/semi-supervised learning. We propose a measure, based on the Odds-Ratio, to quantify how many of these incorrectly classified labels are actually incorrectly labeled and how many of these are confusing. The proposed methods are easily scalable to larger datasets, like ImageNet, LSUN and SUN, as each CNN instance is trained for 60 epochs; or even faster, by implementing a temporal (snapshot) ensemble.

Abstract: In 2015, a picture of a Dress (henceforth the Dress) triggered popular and scientific interest; some reported seeing the Dress in white and gold (W&G) and others in blue and black (B&B). We aimed to describe the phenomenon and investigate the role of contextualization. Few days after the Dress had appeared on the Internet, we projected it to 240 students on two large screens in the classroom. Participants reported seeing the Dress in B&B (48%), W&G (38%), or blue and brown (B&Br; 7%). Amongst numerous socio-demographic variables, we only observed that W&G viewers were most likely to have always seen the Dress as W&G. In the laboratory, we tested how much contextual information is necessary for the phenomenon to occur. Fifty-seven participants selected colours most precisely matching predominant colours of parts or the full Dress. We presented, in this order, small squares (a), vertical strips (b), and the full Dress (c). We found that (1) B&B, B&Br, and W&G viewers had selected colours differing in lightness and chroma levels for contextualized images only (b, c conditions) and hue for fully contextualized condition only (c) and (2) B&B viewers selected colours most closely matching displayed colours of the Dress. Thus, the Dress phenomenon emerges due to inter-individual differences in subjectively perceived lightness, chroma, and hue, at least when all aspects of the picture need to be integrated. Our results support the previous conclusions that contextual information is key to colour perception; it should be important to understand how this actually happens.

Abstract: In this work we introduce a cross modal image retrieval system that allows both text and sketch as input modalities for the query. A cross-modal deep network architecture is formulated to jointly model the sketch and text input modalities as well as the the image output modality, learning a common embedding between text and images and between sketches and images. In addition, an attention model is used to selectively focus the attention on the different objects of the image, allowing for retrieval with multiple objects in the query. Experiments show that the proposed method performs the best in both single and multiple object image retrieval in standard datasets.

Abstract: Libraries are in deep transformation both in EU and around the world, and they are thriving within a great window of opportunity for innovation. In this paper, we show how the Library Living Lab in Barcelona participated of this changing scenario and contributed to create the Bibliolab program, where more than 200 public libraries give voice to their users in a global user-centric innovation initiative, using technology as enabling factor. The Library Living Lab is a real 4-helix implementation where Universities, Research Centers, Public Administration, Companies and the Neighbors are joint together to explore how technology transforms the cultural experience of people. This case is an example of scalability and provides reference tools for policy making, sustainability, user engage methodologies and governance. We provide specific examples of new prototypes and services that help to understand how to redefine the role of the Library as a real hub for social innovation.