A commercial robot, trained by its manufacturer to recognize a predefined number and type of objects, might be used in many settings, that will in general differ in their illumination conditions, background, type and degree of clutter, and so on. Recent computer vision works tackle this generalization issue through domain adaptation methods, assuming as source the visual domain where the system is trained and as target the domain of deployment. All approaches assume to have access to images from all classes of the target during training, an unrealistic condition in robotics applications. We address this issue proposing an algorithm that takes into account the specific needs of robot vision. Our intuition is that the nature of the domain shift experienced mostly in robotics is local. We exploit this through the learning of maps that spatially ground the domain and quantify the degree of shift, embedded into an end-to-end deep domain adaptation architecture. By explicitly localizing the roots of the domain shift we significantly reduce the number of parameters of the architecture to tune, we gain the flexibility necessary to deal with subset of categories in the target domain at training time, and we provide a clear feedback on the rationale behind any classification decision, which can be exploited in human-robot interactions. Experiments on two different settings of the iCub World database confirm the suitability of our method for robot vision.
The effectiveness of generative adversarial approaches in producing images according to a specific style or visual domain has recently opened new directions to solve the unsupervised domain adaptation problem. It has been shown that source labeled images can be modified to mimic target samples making it possible to train directly a classifier in the target domain, despite the original lack of annotated data. Inverse mappings from the target to the source domain have also been evaluated but only passing through adapted feature spaces, thus without new image generation. In this paper we propose to better exploit the potential of generative adversarial networks for adaptation by introducing a novel symmetric mapping among domains. We jointly optimize bi-directional image transformations combining them with target self-labeling. Moreover we define a new class consistency loss that aligns the generators in the two directions imposing to conserve the class identity of an image passing through both domain mappings. A detailed qualitative and quantitative analysis of the reconstructed images confirm the power of our approach. By integrating the two domain specific classifiers obtained with our bi-directional network we exceed previous state-of-the-art unsupervised adaptation results on four different benchmark datasets.
Classifiers trained on given databases perform poorly when tested on data acquired in different settings. This is explained in domain adaptation through a shift among distributions of the source and target domains. Attempts to align them have traditionally resulted in works reducing the domain shift by introducing appropriate loss terms, measuring the discrepancies between source and target distributions, in the objective function. Here we take a different route, proposing to align the learned representations by embedding in any given network specific Domain Alignment Layers, designed to match the source and target feature distributions to a reference one. Opposite to previous works which define a priori in which layers adaptation should be performed, our method is able to automatically learn the degree of feature alignment required at different levels of the deep network. Thorough experiments on different public benchmarks, in the unsupervised setting, confirm the power of our approach.
The instability of myoelectric signals over time complicates their use to control highly articulated prostheses. To address this problem, studies have tried to combine surface electromyography with modalities that are less affected by the amputation and environment, such as accelerometry or gaze information. In the latter case, the hypothesis is that a subject looks at the object he or she intends to manipulate and that knowing this object's affordances allows to constrain the set of possible grasps. In this paper, we develop an automated way to detect stable fixations and show that gaze information is indeed helpful in predicting hand movements. In our multimodal approach, we automatically detect stable gazes and segment an object of interest around the subject's fixation in the visual frame. The patch extracted around this object is subsequently fed through an off-the-shelf deep convolutional neural network to obtain a high level feature representation, which is then combined with traditional surface electromyography in the classification stage. Tests have been performed on a dataset acquired from five intact subjects who performed ten types of grasps on various objects as well as in a functional setting. They show that the addition of gaze information increases the classification accuracy considerably. Further analysis demonstrates that this improvement is consistent for all grasps and concentrated during the movement onset and offset.
Despite the impressive progress brought by deep network in visual object recognition, robot vision is still far from being a solved problem. The most successful convolutional architectures are developed starting from ImageNet, a large scale collection of images of object categories downloaded from the Web. This kind of images is very different from the situated and embodied visual experience of robots deployed in unconstrained settings. To reduce the gap between these two visual experiences, this paper proposes a simple yet effective data augmentation layer that zooms on the object of interest and simulates the object detection outcome of a robot vision system. The layer, that can be used with any convolutional deep architecture, brings to an increase in object recognition performance of up to 7\%, in experiments performed over three different benchmark databases. Upon acceptance of the paper, our robot data augmentation layer will be made publicly available.
This paper presents an approach for semantic place categorization using data obtained from RGB cameras. Previous studies on visual place recognition and classification have shown that, by considering features derived from pre-trained Convolutional Neural Networks (CNNs) in combination with part-based classification models, high recognition accuracy can be achieved, even in presence of occlusions and severe viewpoint changes. Inspired by these works, we propose to exploit local deep representations, representing images as set of regions applying a Na\"{i}ve Bayes Nearest Neighbor (NBNN) model for image classification. As opposed to previous methods where CNNs are merely used as feature extractors, our approach seamlessly integrates the NBNN model into a fully-convolutional neural network. Experimental results show that the proposed algorithm outperforms previous methods based on pre-trained CNN models and that, when employed in challenging robot place recognition tasks, it is robust to occlusions, environmental and sensor changes.
The empirical fact that classifiers, trained on given data collections, perform poorly when tested on data acquired in different settings is theoretically explained in domain adaptation through a shift among distributions of the source and target domains. Alleviating the domain shift problem, especially in the challenging setting where no labeled data are available for the target domain, is paramount for having visual recognition systems working in the wild. As the problem stems from a shift among distributions, intuitively one should try to align them. In the literature, this has resulted in a stream of works attempting to align the feature representations learned from the source and target domains. Here we take a different route. Rather than introducing regularization terms aiming to promote the alignment of the two representations, we act at the distribution level through the introduction of \emph{DomaIn Alignment Layers} (\DIAL), able to match the observed source and target data distributions to a reference one. Thorough experiments on three different public benchmarks we confirm the power of our approach.
Deep networks thrive when trained on large scale data collections. This has given ImageNet a central role in the development of deep architectures for visual object classification. However, ImageNet was created during a specific period in time, and as such it is prone to aging, as well as dataset bias issues. Moving beyond fixed training datasets will lead to more robust visual systems, especially when deployed on robots in new environments which must train on the objects they encounter there. To make this possible, it is important to break free from the need for manual annotators. Recent work has begun to investigate how to use the massive amount of images available on the Web in place of manual image annotations. We contribute to this research thread with two findings: (1) a study correlating a given level of noisily labels to the expected drop in accuracy, for two deep architectures, on two different types of noise, that clearly identifies GoogLeNet as a suitable architecture for learning from Web data; (2) a recipe for the creation of Web datasets with minimal noise and maximum visual variability, based on a visual and natural language processing concept expansion strategy. By combining these two results, we obtain a method for learning powerful deep object models automatically from the Web. We confirm the effectiveness of our approach through object categorization experiments using our Web-derived version of ImageNet on a popular robot vision benchmark database, and on a lifelong object discovery task on a mobile robot.
A number of studies have proposed to use domain adaptation to reduce the training efforts needed to control an upper-limb prosthesis exploiting pre-trained models from prior subjects. These studies generally reported impressive reductions in the required number of training samples to achieve a certain level of accuracy for intact subjects. We further investigate two popular methods in this field to verify whether this result equally applies to amputees. Our findings show instead that this improvement can largely be attributed to a suboptimal hyperparameter configuration. When hyperparameters are appropriately tuned, the standard approach that does not exploit prior information performs on par with the more complicated transfer learning algorithms. Additionally, earlier studies erroneously assumed that the number of training samples relates proportionally to the efforts required from the subject. However, a repetition of a movement is the atomic unit for subjects and the total number of repetitions should therefore be used as reliable measure for training efforts. Also when correcting for this mistake, we do not find any performance increase due to the use of prior models.
Re-identification is generally carried out by encoding the appearance of a subject in terms of outfit, suggesting scenarios where people do not change their attire. In this paper we overcome this restriction, by proposing a framework based on a deep convolutional neural network, SOMAnet, that additionally models other discriminative aspects, namely, structural attributes of the human figure (e.g. height, obesity, gender). Our method is unique in many respects. First, SOMAnet is based on the Inception architecture, departing from the usual siamese framework. This spares expensive data preparation (pairing images across cameras) and allows the understanding of what the network learned. Second, and most notably, the training data consists of a synthetic 100K instance dataset, SOMAset, created by photorealistic human body generation software. Synthetic data represents a good compromise between realistic imagery, usually not required in re-identification since surveillance cameras capture low-resolution silhouettes, and complete control of the samples, which is useful in order to customize the data w.r.t. the surveillance scenario at-hand, e.g. ethnicity. SOMAnet, trained on SOMAset and fine-tuned on recent re-identification benchmarks, outperforms all competitors, matching subjects even with different apparel. The combination of synthetic data with Inception architectures opens up new research avenues in re-identification.