Automatic inference of driving task demand from visual cues of emotion and attention

Sensing the mental, physical and emotional demand of a driving task is of primary importance in road safety research and for effectively designing in-vehicle information systems (IVIS). Particularly, the need of cars capable of sensing and reacting to the emotional state of the driver has been repeatedly advocated in the literature. Algorithms and sensors to identify patterns of human behavior, such as gestures, speech, eye gaze and facial expression, are becoming available by using low cost hardware: This paper presents a new system which uses surrogate measures such as facial expression (emotion) and head pose and movements (intention) to infer task difficulty in a driving situation. 11 drivers were recruited and observed in a simulated driving task that involved several pre-programmed events aimed at eliciting emotive reactions, such as being stuck behind slower vehicles, intersections and roundabouts, and potentially dangerous situations. The resulting system, combining face expressions and head pose classification, is capable of recognizing dangerous events (such as crashes and near misses) and stressful situations (e.g. intersections and way giving) that occur during the simulated drive.

Vision-based simultaneous localization and mapping in changing outdoor environments

For robots operating in outdoor environments, a number of factors, including weather, time of day, rough terrain, high speeds, and hardware limitations, make performing vision-based simultaneous localization and mapping with current techniques infeasible due to factors such as image blur and/or underexposure, especially on smaller platforms and low-cost hardware. In this paper, we present novel visual place-recognition and odometry techniques that address the challenges posed by low lighting, perceptual change, and low-cost cameras. Our primary contribution is a novel two-step algorithm that combines fast low-resolution whole image matching with a higher-resolution patch-verification step, as well as image saliency methods that simultaneously improve performance and decrease computing time. The algorithms are demonstrated using consumer cameras mounted on a small vehicle in a mixed urban and vegetated environment and a car traversing highway and suburban streets, at different times of day and night and in various weather conditions. The algorithms achieve reliable mapping over the course of a day, both when incrementally incorporating new visual scenes from different times of day into an existing map, and when using a static map comprising visual scenes captured at only one point in time. Using the two-step place-recognition process, we demonstrate for the first time single-image, error-free place recognition at recall rates above 50% across a day-night dataset without prior training or utilization of image sequences. This place-recognition performance enables topologically correct mapping across day-night cycles.

Visual place recognition for persistent robot navigation in changing environments

This thesis demonstrates that robots can learn about how the world changes, and can use this information to recognise where they are, even when the appearance of the environment has changed a great deal. The ability to localise in highly dynamic environments using vision only is a key tool for achieving long-term, autonomous navigation in unstructured outdoor environments. The proposed learning algorithms are designed to be unsupervised, and can be generated by the robot online in response to its observations of the world, without requiring information from a human operator or other external source.

On Visual Detection of Highly-occluded Objects for Harvesting Automation in Horticulture

Developing accurate and reliable crop detection algorithms is an important step for harvesting automation in horticulture. This paper presents a novel approach to visual detection of highly-occluded fruits. We use a conditional random field (CRF) on multi-spectral image data (colour and Near-Infrared Reflectance, NIR) to model two classes: crop and background. To describe these two classes, we explore a range of visual-texture features including local binary pattern, histogram of oriented gradients, and learn auto-encoder features. The pro-posed methods are evaluated using hand-labelled images from a dataset captured on a commercial capsicum farm. Experimental results are presented, and performance is evaluated in terms of the Area Under the Curve (AUC) of the precision-recall curves.Our current results achieve a maximum performance of 0.81AUC when combining all of the texture features in conjunction with colour information.

Automatic Lameness Detection in a Milking Robot : Instrumentation, measurement software, algorithms for data analysis and a neural network model

The aim of this thesis is to develop a fully automatic lameness detection system that operates in a milking robot. The instrumentation, measurement software, algorithms for data analysis and a neural network model for lameness detection were developed. Automatic milking has become a common practice in dairy husbandry, and in the year 2006 about 4000 farms worldwide used over 6000 milking robots. There is a worldwide movement with the objective of fully automating every process from feeding to milking. Increase in automation is a consequence of increasing farm sizes, the demand for more efficient production and the growth of labour costs. As the level of automation increases, the time that the cattle keeper uses for monitoring animals often decreases. This has created a need for systems for automatically monitoring the health of farm animals. The popularity of milking robots also offers a new and unique possibility to monitor animals in a single confined space up to four times daily. Lameness is a crucial welfare issue in the modern dairy industry. Limb disorders cause serious welfare, health and economic problems especially in loose housing of cattle. Lameness causes losses in milk production and leads to early culling of animals. These costs could be reduced with early identification and treatment. At present, only a few methods for automatically detecting lameness have been developed, and the most common methods used for lameness detection and assessment are various visual locomotion scoring systems. The problem with locomotion scoring is that it needs experience to be conducted properly, it is labour intensive as an on-farm method and the results are subjective. A four balance system for measuring the leg load distribution of dairy cows during milking in order to detect lameness was developed and set up in the University of Helsinki Research farm Suitia. The leg weights of 73 cows were successfully recorded during almost 10,000 robotic milkings over a period of 5 months. The cows were locomotion scored weekly, and the lame cows were inspected clinically for hoof lesions. Unsuccessful measurements, caused by cows standing outside the balances, were removed from the data with a special algorithm, and the mean leg loads and the number of kicks during milking was calculated. In order to develop an expert system to automatically detect lameness cases, a model was needed. A probabilistic neural network (PNN) classifier model was chosen for the task. The data was divided in two parts and 5,074 measurements from 37 cows were used to train the model. The operation of the model was evaluated for its ability to detect lameness in the validating dataset, which had 4,868 measurements from 36 cows. The model was able to classify 96% of the measurements correctly as sound or lame cows, and 100% of the lameness cases in the validation data were identified. The number of measurements causing false alarms was 1.1%. The developed model has the potential to be used for on-farm decision support and can be used in a real-time lameness monitoring system.

Learning Nonlinear Visual Processing from Natural Images

The paradigm of computational vision hypothesizes that any visual function -- such as the recognition of your grandparent -- can be replicated by computational processing of the visual input. What are these computations that the brain performs? What should or could they be? Working on the latter question, this dissertation takes the statistical approach, where the suitable computations are attempted to be learned from the natural visual data itself. In particular, we empirically study the computational processing that emerges from the statistical properties of the visual world and the constraints and objectives specified for the learning process. This thesis consists of an introduction and 7 peer-reviewed publications, where the purpose of the introduction is to illustrate the area of study to a reader who is not familiar with computational vision research. In the scope of the introduction, we will briefly overview the primary challenges to visual processing, as well as recall some of the current opinions on visual processing in the early visual systems of animals. Next, we describe the methodology we have used in our research, and discuss the presented results. We have included some additional remarks, speculations and conclusions to this discussion that were not featured in the original publications. We present the following results in the publications of this thesis. First, we empirically demonstrate that luminance and contrast are strongly dependent in natural images, contradicting previous theories suggesting that luminance and contrast were processed separately in natural systems due to their independence in the visual data. Second, we show that simple cell -like receptive fields of the primary visual cortex can be learned in the nonlinear contrast domain by maximization of independence. Further, we provide first-time reports of the emergence of conjunctive (corner-detecting) and subtractive (opponent orientation) processing due to nonlinear projection pursuit with simple objective functions related to sparseness and response energy optimization. Then, we show that attempting to extract independent components of nonlinear histogram statistics of a biologically plausible representation leads to projection directions that appear to differentiate between visual contexts. Such processing might be applicable for priming, \ie the selection and tuning of later visual processing. We continue by showing that a different kind of thresholded low-frequency priming can be learned and used to make object detection faster with little loss in accuracy. Finally, we show that in a computational object detection setting, nonlinearly gain-controlled visual features of medium complexity can be acquired sequentially as images are encountered and discarded. We present two online algorithms to perform this feature selection, and propose the idea that for artificial systems, some processing mechanisms could be selectable from the environment without optimizing the mechanisms themselves. In summary, this thesis explores learning visual processing on several levels. The learning can be understood as interplay of input data, model structures, learning objectives, and estimation algorithms. The presented work adds to the growing body of evidence showing that statistical methods can be used to acquire intuitively meaningful visual processing mechanisms. The work also presents some predictions and ideas regarding biological visual processing.

VISUAL OBJECT TRACKING VIA RANDOM FERNS BASED CLASSIFICATION

Designing a robust algorithm for visual object tracking has been a challenging task since many years. There are trackers in the literature that are reasonably accurate for many tracking scenarios but most of them are computationally expensive. This narrows down their applicability as many tracking applications demand real time response. In this paper, we present a tracker based on random ferns. Tracking is posed as a classification problem and classification is done using ferns. We used ferns as they rely on binary features and are extremely fast at both training and classification as compared to other classification algorithms. Our experiments show that the proposed tracker performs well on some of the most challenging tracking datasets and executes much faster than one of the state-of-the-art trackers, without much difference in tracking accuracy.

Algorithms for colour image processing based on neurological models

373 p. : il., gráf., fot., tablas

Skeleton based visual pattern recognition. Applications to tabletop interaction

211 p. :il.

Learning and using taxonomies for visual and olfactory classification

Humans are able of distinguishing more than 5000 visual categories even in complex environments using a variety of different visual systems all working in tandem. We seem to be capable of distinguishing thousands of different odors as well. In the machine learning community, many commonly used multi-class classifiers do not scale well to such large numbers of categories. This thesis demonstrates a method of automatically creating application-specific taxonomies to aid in scaling classification algorithms to more than 100 cate- gories using both visual and olfactory data. The visual data consists of images collected online and pollen slides scanned under a microscope. The olfactory data was acquired by constructing a small portable sniffing apparatus which draws air over 10 carbon black polymer composite sensors. We investigate performance when classifying 256 visual categories, 8 or more species of pollen and 130 olfactory categories sampled from common household items and a standardized scratch-and-sniff test. Taxonomies are employed in a divide-and-conquer classification framework which improves classification time while allowing the end user to trade performance for specificity as needed. Before classification can even take place, the pollen counter and electronic nose must filter out a high volume of background “clutter” to detect the categories of interest. In the case of pollen this is done with an efficient cascade of classifiers that rule out most non-pollen before invoking slower multi-class classifiers. In the case of the electronic nose, much of the extraneous noise encountered in outdoor environments can be filtered using a sniffing strategy which preferentially samples the visensor response at frequencies that are relatively immune to background contributions from ambient water vapor. This combination of efficient background rejection with scalable classification algorithms is tested in detail for three separate projects: 1) the Caltech-256 Image Dataset, 2) the Caltech Automated Pollen Identification and Counting System (CAPICS) and 3) a portable electronic nose specially constructed for outdoor use.

Color to Gray: Visual Cue Preservation

Both commercial and scientific applications often need to transform color images into gray-scale images, e. g., to reduce the publication cost in printing color images or to help color blind people see visual cues of color images. However, conventional color to gray algorithms are not ready for practical applications because they encounter the following problems: 1) Visual cues are not well defined so it is unclear how to preserve important cues in the transformed gray-scale images; 2) some algorithms have extremely high time cost for computation; and 3) some require human-computer interactions to have a reasonable transformation. To solve or at least reduce these problems, we propose a new algorithm based on a probabilistic graphical model with the assumption that the image is defined over a Markov random field. Thus, color to gray procedure can be regarded as a labeling process to preserve the newly well-defined visual cues of a color image in the transformed gray-scale image. Visual cues are measurements that can be extracted from a color image by a perceiver. They indicate the state of some properties of the image that the perceiver is interested in perceiving. Different people may perceive different cues from the same color image and three cues are defined in this paper, namely, color spatial consistency, image structure information, and color channel perception priority. We cast color to gray as a visual cue preservation procedure based on a probabilistic graphical model and optimize the model based on an integral minimization problem. We apply the new algorithm to both natural color images and artificial pictures, and demonstrate that the proposed approach outperforms representative conventional algorithms in terms of effectiveness and efficiency. In addition, it requires no human-computer interactions.

Visual Navigation: Constructing and Utilizing Simple Maps of an Indoor Environment

The goal of this work is to navigate through an office environmentsusing only visual information gathered from four cameras placed onboard a mobile robot. The method is insensitive to physical changes within the room it is inspecting, such as moving objects. Forward and rotational motion vision are used to find doors and rooms, and these can be used to build topological maps. The map is built without the use of odometry or trajectory integration. The long term goal of the project described here is for the robot to build simple maps of its environment and to localize itself within this framework.

Texture Segregation By Visual Cortex: Perceptual Grouping, Attention, and Learning

A neural model is proposed of how laminar interactions in the visual cortex may learn and recognize object texture and form boundaries. The model brings together five interacting processes: region-based texture classification, contour-based boundary grouping, surface filling-in, spatial attention, and object attention. The model shows how form boundaries can determine regions in which surface filling-in occurs; how surface filling-in interacts with spatial attention to generate a form-fitting distribution of spatial attention, or attentional shroud; how the strongest shroud can inhibit weaker shrouds; and how the winning shroud regulates learning of texture categories, and thus the allocation of object attention. The model can discriminate abutted textures with blurred boundaries and is sensitive to texture boundary attributes like discontinuities in orientation and texture flow curvature as well as to relative orientations of texture elements. The model quantitatively fits a large set of human psychophysical data on orientation-based textures. Object boundar output of the model is compared to computer vision algorithms using a set of human segmented photographic images. The model classifies textures and suppresses noise using a multiple scale oriented filterbank and a distributed Adaptive Resonance Theory (dART) classifier. The matched signal between the bottom-up texture inputs and top-down learned texture categories is utilized by oriented competitive and cooperative grouping processes to generate texture boundaries that control surface filling-in and spatial attention. Topdown modulatory attentional feedback from boundary and surface representations to early filtering stages results in enhanced texture boundaries and more efficient learning of texture within attended surface regions. Surface-based attention also provides a self-supervising training signal for learning new textures. Importance of the surface-based attentional feedback in texture learning and classification is tested using a set of textured images from the Brodatz micro-texture album. Benchmark studies vary from 95.1% to 98.6% with attention, and from 90.6% to 93.2% without attention.

Different stimuli, different spatial codes: a visual map and an auditory rate code for oculomotor space in the primate superior colliculus.

Maps are a mainstay of visual, somatosensory, and motor coding in many species. However, auditory maps of space have not been reported in the primate brain. Instead, recent studies have suggested that sound location may be encoded via broadly responsive neurons whose firing rates vary roughly proportionately with sound azimuth. Within frontal space, maps and such rate codes involve different response patterns at the level of individual neurons. Maps consist of neurons exhibiting circumscribed receptive fields, whereas rate codes involve open-ended response patterns that peak in the periphery. This coding format discrepancy therefore poses a potential problem for brain regions responsible for representing both visual and auditory information. Here, we investigated the coding of auditory space in the primate superior colliculus(SC), a structure known to contain visual and oculomotor maps for guiding saccades. We report that, for visual stimuli, neurons showed circumscribed receptive fields consistent with a map, but for auditory stimuli, they had open-ended response patterns consistent with a rate or level-of-activity code for location. The discrepant response patterns were not segregated into different neural populations but occurred in the same neurons. We show that a read-out algorithm in which the site and level of SC activity both contribute to the computation of stimulus location is successful at evaluating the discrepant visual and auditory codes, and can account for subtle but systematic differences in the accuracy of auditory compared to visual saccades. This suggests that a given population of neurons can use different codes to support appropriate multimodal behavior.