The role of context in image annotation and recommendation

With the rise of smart phones, lifelogging devices (e.g. Google Glass) and popularity of image sharing websites (e.g. Flickr), users are capturing and sharing every aspect of their life online producing a wealth of visual content. Of these uploaded images, the majority are poorly annotated or exist in complete semantic isolation making the process of building retrieval systems difficult as one must firstly understand the meaning of an image in order to retrieve it. To alleviate this problem, many image sharing websites offer manual annotation tools which allow the user to “tag” their photos, however, these techniques are laborious and as a result have been poorly adopted; Sigurbjörnsson and van Zwol (2008) showed that 64% of images uploaded to Flickr are annotated with < 4 tags. Due to this, an entire body of research has focused on the automatic annotation of images (Hanbury, 2008; Smeulders et al., 2000; Zhang et al., 2012a) where one attempts to bridge the semantic gap between an image’s appearance and meaning e.g. the objects present. Despite two decades of research the semantic gap still largely exists and as a result automatic annotation models often offer unsatisfactory performance for industrial implementation. Further, these techniques can only annotate what they see, thus ignoring the “bigger picture” surrounding an image (e.g. its location, the event, the people present etc). Much work has therefore focused on building photo tag recommendation (PTR) methods which aid the user in the annotation process by suggesting tags related to those already present. These works have mainly focused on computing relationships between tags based on historical images e.g. that NY and timessquare co-exist in many images and are therefore highly correlated. However, tags are inherently noisy, sparse and ill-defined often resulting in poor PTR accuracy e.g. does NY refer to New York or New Year? This thesis proposes the exploitation of an image’s context which, unlike textual evidences, is always present, in order to alleviate this ambiguity in the tag recommendation process. Specifically we exploit the “what, who, where, when and how” of the image capture process in order to complement textual evidences in various photo tag recommendation and retrieval scenarios. In part II, we combine text, content-based (e.g. # of faces present) and contextual (e.g. day-of-the-week taken) signals for tag recommendation purposes, achieving up to a 75% improvement to precision@5 in comparison to a text-only TF-IDF baseline. We then consider external knowledge sources (i.e. Wikipedia & Twitter) as an alternative to (slower moving) Flickr in order to build recommendation models on, showing that similar accuracy could be achieved on these faster moving, yet entirely textual, datasets. In part II, we also highlight the merits of diversifying tag recommendation lists before discussing at length various problems with existing automatic image annotation and photo tag recommendation evaluation collections. In part III, we propose three new image retrieval scenarios, namely “visual event summarisation”, “image popularity prediction” and “lifelog summarisation”. In the first scenario, we attempt to produce a rank of relevant and diverse images for various news events by (i) removing irrelevant images such memes and visual duplicates (ii) before semantically clustering images based on the tweets in which they were originally posted. Using this approach, we were able to achieve over 50% precision for images in the top 5 ranks. In the second retrieval scenario, we show that by combining contextual and content-based features from images, we are able to predict if it will become “popular” (or not) with 74% accuracy, using an SVM classifier. Finally, in chapter 9 we employ blur detection and perceptual-hash clustering in order to remove noisy images from lifelogs, before combining visual and geo-temporal signals in order to capture a user’s “key moments” within their day. We believe that the results of this thesis show an important step towards building effective image retrieval models when there lacks sufficient textual content (i.e. a cold start).

Machine learning techniques for implicit interaction using mobile sensors

Interactions in mobile devices normally happen in an explicit manner, which means that they are initiated by the users. Yet, users are typically unaware that they also interact implicitly with their devices. For instance, our hand pose changes naturally when we type text messages. Whilst the touchscreen captures finger touches, hand movements during this interaction however are unused. If this implicit hand movement is observed, it can be used as additional information to support or to enhance the users’ text entry experience. This thesis investigates how implicit sensing can be used to improve existing, standard interaction technique qualities. In particular, this thesis looks into enhancing front-of-device interaction through back-of-device and hand movement implicit sensing. We propose the investigation through machine learning techniques. We look into problems on how sensor data via implicit sensing can be used to predict a certain aspect of an interaction. For instance, one of the questions that this thesis attempts to answer is whether hand movement during a touch targeting task correlates with the touch position. This is a complex relationship to understand but can be best explained through machine learning. Using machine learning as a tool, such correlation can be measured, quantified, understood and used to make predictions on future touch position. Furthermore, this thesis also evaluates the predictive power of the sensor data. We show this through a number of studies. In Chapter 5 we show that probabilistic modelling of sensor inputs and recorded touch locations can be used to predict the general area of future touches on touchscreen. In Chapter 7, using SVM classifiers, we show that data from implicit sensing from general mobile interactions is user-specific. This can be used to identify users implicitly. In Chapter 6, we also show that touch interaction errors can be detected from sensor data. In our experiment, we show that there are sufficient distinguishable patterns between normal interaction signals and signals that are strongly correlated with interaction error. In all studies, we show that performance gain can be achieved by combining sensor inputs.