Doctor of Philosophy / Department of Computer Science / David A. Gustafson / In this paper we evaluate a general form of image classification algorithm based on dense SIFT sampling. This algorithm is present in some form in most state-of-the-art classification systems. However, in this algorithm, numerous parameters must be tuned, and current research provides little insight into effective parameter tuning. We explore the relationship between various parameters and classification performance. Many of our results suggest that there are basic modifications which would improve state-of-the-art algorithms. Additionally, we develop two novel concepts, sampling redundancy and semantic capacity, to explain our data. These concepts provide additional insight into the limitations and potential improvements of state-of-the-art algorithms.

Dans le cadre de cette étude, nous proposons d’analyser la langue militaire et particulièrementcelle de l’armée de Terre. Dans un premier temps, nous traitons de la terminologie, de lalangue militaire technique par rapport à la langue générale et d’un aperçu des dictionnaires despécialité et glossaires de la Défense. Dans une deuxième partie, nous nous appuyons sur lanotion des classes d’objets, instaurées par G. GROSS, qui demeure la meilleure méthoded’analyse de la langue militaire en étudiant les propriétés syntaxiques des prédicats nominauxd’action, les verbes supports et appropriés et leur actualisation. Pour illustrer et approfondirnotre travail, nous nous basons sur les actions des militaires en préparation d’une opération,pendant la projection et à l’issue de l’engagement. Par ces actions, nous démontrons toutes lespropriétés syntaxiques et sémantiques spécifiques à chaque classe de prédicats. La troisièmepartie, consacrée aux arguments de type humain, inanimé concret et locatif, est la suitelogique de l’étude des classes d’objets. Avec cette thèse, nous possédons tous les élémentspour envisager, par exemple une étude contrastive avec une autre langue et surtout grâce auxanalyses, la création d’un dictionnaire bilingue. / Within the framework of this study, we propose to analyse military language and particularlythat of the army. In part One, we deal with terminology, technical military language inrelation to everyday language as well as a survey of Defense-related specialized dictionariesand glossaries. In part Two, we take our cue from G. GROSS’s concept of object classes,which remains the best way of approaching military language specifically action-orientednominal predicates’syntactical properties, and support and appropriate verbs. We drew ourexamples from military action during fieldwork, both home and abroad, and upon the returnhome. Making use of these three phases, we outline syntactical and semantic propertiesspecific to each class of predicates. Part Three is devoted to arguments of three types : human,inanimate and locative ; it follows logically upon our analysis of object classes. Thisdissertation will help us carry out a contrastive analysis with another language as well as thecreation of a bilingual dictionary.

A central task in computer vision is detecting object classes such as cars and horses in complex scenes. Training an object class detector typically requires a large set of images labeled with tight bounding boxes around every object instance. Obtaining such data requires human annotation, which is very expensive and time consuming. Alternatively, researchers have tried to train models in a weakly supervised setting (i.e., given only image-level labels), which is much cheaper but leads to weaker detectors. In this thesis, we propose new and efficient human annotation schemes for training object class detectors that bypass the need for drawing bounding boxes and reduce the annotation cost while still obtaining high quality object detectors. First, we propose to train object class detectors from eye tracking data. Instead of drawing tight bounding boxes, the annotators only need to look at the image and find the target object. We track the eye movements of annotators while they perform this visual search task and we propose a technique for deriving object bounding boxes from these eye fixations. To validate our idea, we augment an existing object detection dataset with eye tracking data. Second, we propose a scheme for training object class detectors, which only requires annotators to verify bounding-boxes produced automatically by the learning algorithm. Our scheme introduces human verification as a new step into a standard weakly supervised framework which typically iterates between re-training object detectors and re-localizing objects in the training images. We use the verification signal to improve both re-training and re-localization. Third, we propose another scheme where annotators are asked to click on the center of an imaginary bounding box, which tightly encloses the object. We then incorporate these clicks into a weakly supervised object localization technique, to jointly localize object bounding boxes over all training images. Both our center-clicking and human verification schemes deliver detectors performing almost as well as those trained in a fully supervised setting. Finally, we propose extreme clicking. We ask the annotator to click on four physical points on the object: the top, bottom, left- and right-most points. This task is more natural than the traditional way of drawing boxes and these points are easy to find. Our experiments show that annotating objects with extreme clicking is 5 X faster than the traditional way of drawing boxes and it leads to boxes of the same quality as the original ground-truth drawn the traditional way. Moreover, we use the resulting extreme points to obtain more accurate segmentations than those derived from bounding boxes.