Adrian Muscat

Automatic description generation from images: a survey of models, datasets, and evaluation measures

Automatic description generation from natural images is a challenging problem that has recently received a large amount of interest from the computer vision and natural language processing communities. In this survey, we classify the existing approaches based on how they conceptualize this problem, viz., models that cast description as either generation problem or as a retrieval problem over a visual or multimodal representational space. We provide a detailed review of existing models, highlighting their advantages and disadvantages. Moreover, we give an overview of the benchmark image datasets and the evaluation measures that have been developed to assess the quality of machine-generated image descriptions. Finally we extrapolate future directions in the area of automatic image description generation.

Tunable microstrip antenna using switchable patches

The latest and next generation mobile radio terminals require access to different wireless networks that operate on different radio frequency bands. Additionally requirements such as a varying asymmetric data rate are needed for the efficient transmission of multimedia services. To operate effectively such terminals need either a single ultra wideband antenna that spans all the useful mobile frequencies, or an antenna that can be tuned to the required relatively narrow bands when desirable. In this paper a low-profile tunable microstrip antenna structure using switched parasitic patches is studied. The frequency of operation of the prototype can be tuned from 1.36 GHz to 1.92 GHz using a total of 13 different switch combinations. The antenna is simulated, built and tested. The switches in the prototype are replaced with copper strip bridges. The measured results are in good agreement with the simulations.

Modeling electromagnetic interference generated by a WLAN system onboard commercial aircraft

The growing demand for the utilization of personal electronic communication devices onboard commercial aircraft necessitates the assurance of safety by airline operators and regulators. This implies that the potential risks posed by the deployment of wireless communication systems on critical aircraft equipment must be carefully assessed and countermeasures taken when required. In this paper, a model based on a ray-tracing algorithm is developed to calculate the electromagnetic interference incident on the fuselage structure of a commercial airline. The source of interference is a 2.4 GHz data communications network. Two scenarios are considered; the first assumes a base station in the centre of the cabin while the second considers four base stations, transmitting at a lower power, distributed along the cabin. The model first determines the propagation map generated by the base stations. These results are used to establish the transmission power required by the personal mobile devices which is then employed to determine the propagation map of each device. The overall electromagnetic interference map incident on the fuselage resulting from the onboard wireless network is generated by vectorially combining the resulting

Mobile gaming on a virtualized infrastructure

This paper describes the design, implementation and testing of a prototype system that processes and manages games on a remote server and played on a mobile device. The game application is executed on the remote server and video-streamed to a mobile device over WiFi. The mobile keyboard input is relayed to the server to control the game. The prototype test-bed includes a virtualization based infrastructure and the necessary software for both the virtual game server and the mobile device. Performance tests are carried out on this prototype test-bed and the results are evaluated.

A small tunable antenna using multiple shorting posts and varactor diodes

An electrically small tunable antenna for use in mobile terminals is designed. The small size of the antenna is achieved by the use of multiple shorting posts and varactor diodes. Multiple shorting posts are used to allow a greater freedom in the placement of the probe feed. Two varactor diodes placed in line with the shorting posts are used to further decrease the size of the antenna. The antenna can be tuned over a range from 0.55GHz to 1.5GHz. However the antenna has a small gain at the lower frequencies due to its small size.

Generating Descriptions of Spatial Relations between Objects in Images

We investigate the task of predicting prepositions that can be used to describe the spatial relationships between pairs of objects depicted in images. We explore the extent to which such spatial prepositions can be predicted from (a) language information, (b) visual information, and (c) combinations of the two. In this paper we describe the dataset of object pairs and prepositions we have created, and report first results for predicting prepositions for object pairs, using a Naive Bayes framework. The features we use include object class labels and geometrical features computed from object bounding boxes. We evaluate the results in terms of accuracy against human-selected prepositions.

Describing Spatial Relationships between Objects in Images in English and French

The context for the work we report here is the automatic description of spatial relationships between pairs of objects in images. We investigate the task of selecting prepositions for such spatial relationships. We describe the two datasets of object pairs and prepositions we have created for English and French, and report results for predicting prepositions for object pairs in both of these languages, using two methods: (a) an existing approach which manually fixes the mapping from geometrical features to prepositions, and (b) a Naive Bayes classifier trained on the English and French datasets. For the latter we use features based on object class labels and geometrical measurements of object bounding boxes. We evaluate the automatically generated prepositions on unseen data in terms of accuracy against the human-selected prepositions.

An analytical method of assessment of RemoteFX as a cloud gaming platform

This paper describes a cloud computing platform that executes graphics intensive programs, such as computer games, with the support of a GPU to render the graphics, and stream the ensuing video to a mobile device over bandwidth constrained channels. Controls on the mobile device allow the user to interact with the game remotely. Initial results based on a Microsoft Windows platform using the Hypervisor in Windows Server 2012, with RemoteFX are reported.

Objective video quality metrics for HDTV services: A survey

The exponential growth of video traffic is expected to reach 62% of the global Internet traffic by the end of 2015 [1]. This presents as a significant challenge for the television service providers who need to employ networking technologies to monitor specific Quality of Service (QoS) parameters such as packet loss rate, jitter and delay, to ensure an acceptable level of quality. However, recent research has demonstrated that the quality experienced by the end-user does not correlate to the QoS parameters employed by most service providers [2]. This paper investigates the correlation between the QoS parameters and the quality perceived by the end. user. These results indicate that although the QoS parameters may sometimes achieve high correlation with respect to the quality perceived by the viewer, they still have large variances. This suggests that the QoS parameters are not enough to quantify the subjective quality with a high level of confidence. This work further compares a number of existing objective video quality metrics. The results presented in this paper show that the Full-Reference Motion based Video Integrity Evaluation (MOVIE) metric and the Spatio-Temporal Reduced Reference Entropic Differences (STRRED) metric achieve excellent correlation with the subjective scores. This research also demonstrates that the STRRED metric and its derivatives have several advantages over the MOVIE metric since less information needs to be transmitted and it is less computationally intensive.

Small aircraft cockpit electromagnetic interference due to UMTS signal propagation

This paper implemented a ray-tracing based algorithm to calculate the EMI generated inside an aircraft cabin with particular emphasis on flight critical equipment such as the cockpit avionics panel. The interference level caused by the deployment of a UMTS Node B antenna for the provision of passenger communication was computed. The model was employed to study the effects of the multipath electromagnetic propagation attained from positioning a Node B antenna in different locations inside a relatively small tunnel shaped fuselage with dense furnishings. The proposed framework is also able to provide an all-encompassing understanding of the generated EMI, whilst offering support for the development of conscientious network design in aircraft.