Mid-level concept learning with visual contextual ontologies and probabilistic inference for image annotation

To date, automatic recognition of semantic information such as salient objects and mid-level concepts from images is a challenging task. Since real-world objects tend to exist in a context within their environment, the computer vision researchers have increasingly incorporated contextual information for improving object recognition. In this paper, we present a method to build a visual contextual ontology from salient objects descriptions for image annotation. The ontologies include not only partOf/kindOf relations, but also spatial and co-occurrence relations. A two-step image annotation algorithm is also proposed based on ontology relations and probabilistic inference. Different from most of the existing work, we specially exploit how to combine representation of ontology, contextual knowledge and probabilistic inference. The experiments show that image annotation results are improved in the LabelMe dataset.

Ontology-based image annotation

With regard to the long-standing problem of the semantic gap between low-level image features and high-level human knowledge, the image retrieval community has recently shifted its emphasis from low-level features analysis to high-level image semantics extrac- tion. User studies reveal that users tend to seek information using high-level semantics. Therefore, image semantics extraction is of great importance to content-based image retrieval because it allows the users to freely express what images they want. Semantic content annotation is the basis for semantic content retrieval. The aim of image anno- tation is to automatically obtain keywords that can be used to represent the content of images. The major research challenges in image semantic annotation are: what is the basic unit of semantic representation? how can the semantic unit be linked to high-level image knowledge? how can the contextual information be stored and utilized for image annotation? In this thesis, the Semantic Web technology (i.e. ontology) is introduced to the image semantic annotation problem. Semantic Web, the next generation web, aims at mak- ing the content of whatever type of media not only understandable to humans but also to machines. Due to the large amounts of multimedia data prevalent on the Web, re- searchers and industries are beginning to pay more attention to the Multimedia Semantic Web. The Semantic Web technology provides a new opportunity for multimedia-based applications, but the research in this area is still in its infancy. Whether ontology can be used to improve image annotation and how to best use ontology in semantic repre- sentation and extraction is still a worth-while investigation. This thesis deals with the problem of image semantic annotation using ontology and machine learning techniques in four phases as below. 1) Salient object extraction. A salient object servers as the basic unit in image semantic extraction as it captures the common visual property of the objects. Image segmen- tation is often used as the �rst step for detecting salient objects, but most segmenta- tion algorithms often fail to generate meaningful regions due to over-segmentation and under-segmentation. We develop a new salient object detection algorithm by combining multiple homogeneity criteria in a region merging framework. 2) Ontology construction. Since real-world objects tend to exist in a context within their environment, contextual information has been increasingly used for improving object recognition. In the ontology construction phase, visual-contextual ontologies are built from a large set of fully segmented and annotated images. The ontologies are composed of several types of concepts (i.e. mid-level and high-level concepts), and domain contextual knowledge. The visual-contextual ontologies stand as a user-friendly interface between low-level features and high-level concepts. 3) Image objects annotation. In this phase, each object is labelled with a mid-level concept in ontologies. First, a set of candidate labels are obtained by training Support Vectors Machines with features extracted from salient objects. After that, contextual knowledge contained in ontologies is used to obtain the �nal labels by removing the ambiguity concepts. 4) Scene semantic annotation. The scene semantic extraction phase is to get the scene type by using both mid-level concepts and domain contextual knowledge in ontologies. Domain contextual knowledge is used to create scene con�guration that describes which objects co-exist with which scene type more frequently. The scene con�guration is represented in a probabilistic graph model, and probabilistic inference is employed to calculate the scene type given an annotated image. To evaluate the proposed methods, a series of experiments have been conducted in a large set of fully annotated outdoor scene images. These include a subset of the Corel database, a subset of the LabelMe dataset, the evaluation dataset of localized semantics in images, the spatial context evaluation dataset, and the segmented and annotated IAPR TC-12 benchmark.

Automatic Image Annotation Using SURF Descriptors

In recent years there is an apparent shift in research from content based image retrieval (CBIR) to automatic image annotation in order to bridge the gap between low level features and high level semantics of images. Automatic Image Annotation (AIA) techniques facilitate extraction of high level semantic concepts from images by machine learning techniques. Many AIA techniques use feature analysis as the first step to identify the objects in the image. However, the high dimensional image features make the performance of the system worse. This paper describes and evaluates an automatic image annotation framework which uses SURF descriptors to select right number of features and right features for annotation. The proposed framework uses a hybrid approach in which k-means clustering is used in the training phase and fuzzy K-NN classification in the annotation phase. The performance of the system is evaluated using standard metrics.

Clustering based multi-label classification for image annotation and retrieval

This paper presents a novel multi-label classification framework for domains with large numbers of labels. Automatic image annotation is such a domain, as the available semantic concepts are typically hundreds. The proposed framework comprises an initial clustering phase that breaks the original training set into several disjoint clusters of data. It then trains a multi-label classifier from the data of each cluster. Given a new test instance, the framework first finds the nearest cluster and then applies the corresponding model. Empirical results using two clustering algorithms, four multi-label classification algorithms and three image annotation data sets suggest that the proposed approach can improve the performance and reduce the training time of standard multi-label classification algorithms, particularly in the case of large number of labels.

Empirical study of multi-label classification methods for image annotation and retrieval

This paper presents an empirical study of multi-label classification methods, and gives suggestions for multi-label classification that are effective for automatic image annotation applications. The study shows that triple random ensemble multi-label classification algorithm (TREMLC) outperforms among its counterparts, especially on scene image dataset. Multi-label k-nearest neighbor (ML-kNN) and binary relevance (BR) learning algorithms perform well on Corel image dataset. Based on the overall evaluation results, examples are given to show label prediction performance for the algorithms using selected image examples. This provides an indication of the suitability of different multi-label classification methods for automatic image annotation under different problem settings.

Image annotation with Photocopain

Photo annotation is a resource-intensive task, yet is increasingly essential as image archives and personal photo collections grow in size. There is an inherent con?ict in the process of describing and archiving personal experiences, because casual users are generally unwilling to expend large amounts of e?ort on creating the annotations which are required to organise their collections so that they can make best use of them. This paper describes the Photocopain system, a semi-automatic image annotation system which combines information about the context in which a photograph was captured with information from other readily available sources in order to generate outline annotations for that photograph that the user may further extend or amend.

Task-based image annotation and retrieval

In order to address problems of information overload in digital imagery task domains we have developed an interactive approach to the capture and reuse of image context information. Our framework models different aspects of the relationship between images and domain tasks they support by monitoring the interactive manipulation and annotation of task-relevant imagery. The approach allows us to gauge a measure of a user's intentions as they complete goal-directed image tasks. As users analyze retrieved imagery their interactions are captured and an expert task context is dynamically constructed. This human expertise, proficiency, and knowledge can then be leveraged to support other users in carrying out similar domain tasks. We have applied our techniques to two multimedia retrieval applications for two different image domains, namely the geo-spatial and medical imagery domains. © Springer-Verlag Berlin Heidelberg 2007.

Deep Corals, Deep Learning: Moving the Deep Net Towards Real-Time Image Annotation

The mismatch between human capacity and the acquisition of Big Data such as Earth imagery undermines commitments to Convention on Biological Diversity (CBD) and Aichi targets. Artificial intelligence (AI) solutions to Big Data issues are urgently needed as these could prove to be faster, more accurate, and cheaper. Reducing costs of managing protected areas in remote deep waters and in the High Seas is of great importance, and this is a realm where autonomous technology will be transformative.

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).

High-level concept annotation using ontology and probabilistic inference

Image annotation is a significant step towards semantic based image retrieval. Ontology is a popular approach for semantic representation and has been intensively studied for multimedia analysis. However, relations among concepts are seldom used to extract higher-level semantics. Moreover, the ontology inference is often crisp. This paper aims to enable sophisticated semantic querying of images, and thus contributes to 1) an ontology framework to contain both visual and contextual knowledge, and 2) a probabilistic inference approach to reason the high-level concepts based on different sources of information. The experiment on a natural scene database from LabelMe database shows encouraging results.

Image to text translation by multi-label classification

This paper presents an image to text translation platform consisting of image segmentation, region features extraction, region blobs clustering, and translation components. A multi-label learning method is suggested for realizing the translation component. Empirical studies show that the predictive performance of the translation component is better than its counterparts when employed a dual-random ensemble multi-label classification algorithm.

ImageSpace: An Environment for Image Ontology Management

More and more researchers have realized that ontologies will play a critical role in the development of the Semantic Web, the next generation Web in which content is not only consumable by humans, but also by software agents. The development of tools to support ontology management including creation, visualization, annotation, database storage, and retrieval is thus extremely important. We have developed ImageSpace, an image ontology creation and annotation tool that features (1) full support for the standard web ontology language DAML+OIL; (2) image ontology creation, visualization, image annotation and display in one integrated framework; (3) ontology consistency assurance; and (4) storing ontologies and annotations in relational databases. It is expected that the availability of such a tool will greatly facilitate the creation of image repositories as islands of the Semantic Web.

Can publicly available webcameras and mechanical Turks be used to evaluate physical activity policy and built environment change?

Introduction: Fewer than 50% of adults and 40% of youth meet US CDC guidelines for physical activity (PA) with the built environment (BE) a culprit for limited PA. A challenge in evaluating policy and BE change is the forethought to capture a priori PA behaviors and the ability to eliminate bias in post-change environments. The present objective was to analyze existing public data feeds to quantify effectiveness of BE interventions. The Archive of Many Outdoor Scenes (AMOS) has collected 135 million images of outdoor environments from 12,000 webcams since 2006. Many of these environments have experienced BE change. Methods: One example of BE change is the addition of protected bike lanes and a bike share program in Washington, DC.Weselected an AMOS webcam that captured this change. AMOS captures a photograph from eachwebcamevery half hour.AMOScaptured the 120 webcam photographs between 0700 and 1900 during the first work week of June 2009 and the 120 photographs from the same week in 2010. We used the Amazon Mechanical Turk (MTurk) website to crowd-source the image annotation. MTurk workers were paid US$0.01 to mark each pedestrian, cyclist and vehicle in a photograph. Each image was coded 5 unique times (n=1200). The data, counts of transportation mode, was downloaded to SPSS for analysis. Results: The number of cyclists per scene increased four-fold between 2009 and 2010 (F=36.72, p=0.002). There was no significant increase in pedestrians between the two years, however there was a significant increase in number of vehicles per scene (F=16.81, p