Assessment of crowdsourcing and gamification loss in user-assisted object segmentation

There has been a growing interest in applying human computation – particularly crowdsourcing techniques – to assist in the solution of multimedia, image processing, and computer vision problems which are still too difficult to solve using fully automatic algorithms, and yet relatively easy for humans. In this paper we focus on a specific problem – object segmentation within color images – and compare different solutions which combine color image segmentation algorithms with human efforts, either in the form of an explicit interactive segmentation task or through an implicit collection of valuable human traces with a game. We use Click’n’Cut, a friendly, web-based, interactive segmentation tool that allows segmentation tasks to be assigned to many users, and Ask’nSeek, a game with a purpose designed for object detection and segmentation. The two main contributions of this paper are: (i) We use the results of Click’n’Cut campaigns with different groups of users to examine and quantify the crowdsourcing loss incurred when an interactive segmentation task is assigned to paid crowd-workers, comparing their results to the ones obtained when computer vision experts are asked to perform the same tasks. (ii) Since interactive segmentation tasks are inherently tedious and prone to fatigue, we compare the quality of the results obtained with Click’n’Cut with the ones obtained using a (fun, interactive, and potentially less tedious) game designed for the same purpose. We call this contribution the assessment of the gamification loss, since it refers to how much quality of segmentation results may be lost when we switch to a game-based approach to the same task. We demonstrate that the crowdsourcing loss is significant when using all the data points from workers, but decreases substantially (and becomes comparable to the quality of expert users performing similar tasks) after performing a modest amount of data analysis and filtering out of users whose data are clearly not useful. We also show that – on the other hand – the gamification loss is significantly more severe: the quality of the results drops roughly by half when switching from a focused (yet tedious) task to a more fun and relaxed game environment.     

Neural network approach to background modeling for video object segmentation.

This paper presents a novel background modeling and subtraction approach for video object segmentation. A neural network (NN) architecture is proposed to form an unsupervised Bayesian classifier for this application domain. The constructed classifier efficiently handles the segmentation in natural-scene sequences with complex background motion and changes in illumination. The weights of the proposed NN serve as a model of the background and are temporally updated to reflect the observed statistics of background. The segmentation performance of the proposed NN is qualitatively and quantitatively examined and compared to two extant probabilistic object segmentation algorithms, based on a previously published test pool containing diverse surveillance-related sequences. The proposed algorithm is parallelized on a subpixel level and designed to enable efficient hardware implementation.     

Traumatic subdural pneumocephalus causing rise in intracranial pressure in the early phase of head trauma: report of two cases

The mortality rate is high and prognosis is worse among new-borns with prenatal diagnosis of heart malformation, mainly due to factors such as its association with other malformations, and a range of more severe diseases probably resulting from the predominance of the obstetric use of the four chamber view. In this study we retrospectively assessed the range of cardiopathies diagnosed by foetal echocardiography and their evolution, compared with previous years. From January 1994 to December 1995, 1173 foetal echocardiograms were performed at a gestation age of 24 weeks. Sixty-one foetuses (5.2%) had cardiac anomalies, structural in 56 and arrhythmia in 5. The risks and indications were maternal in 37%, foetal in 31%, familial in 17% and environmental in 15%. Three were false negatives (VSD:2; truncus arteriosus: 1). Five died in utero, and 18 were assessed after birth with a mean gestational age of 37 weeks and birth weight of 3 Kg, a caesarean section was performed in 9. All but one were born in central hospitals. Six children were operated on. Two children died, one after surgery. Compared with the four previous years of activity, indication due to foetal risk rose from 6 to 31%, the number of cases diagnosed with heart disease increased from 14 to 30 per year, and the mortality decreased from 59 to 11%. Despite this, we still observe that the vast majority of new-borns who are hospitalised due to a severe heart disease had no prenatal diagnosis, indicating the need to continue our educational policy in this field.     

A novel neural network based approach to latent overlapped fingerprints separation

Multimedia Tools and Applications - Overlapped fingerprints are often found in latent fingerprints lifted from crime scenes and in live-scan fingerprint images when the surface of fingerprint...     

MUSE: A Content-Based Image Search and Retrieval System Using Relevance Feedback

The field of Content-Based Visual Information Retrieval (CBVIR) has experienced tremendous growth in the recent years and many research groups are currently working on solutions to the problem of finding a desired image or video clip in a huge archive without resorting to metadata. This paper describes the ongoing development of a CBVIR system for image search and retrieval with relevance feedback capabilities. It supports browsing, query-by-example, and two different relevance feedback modes that allow users to refine their queries by indicating which images are good or bad at each iteration.     

A novel tool for summarization of arthroscopic videos

Arthroscopic surgery is a minimally invasive procedure that uses a small camera to generate video streams, which are recorded and subsequently archived. In this paper we present a video summarization tool and demonstrate how it can be successfully used in the domain of arthroscopic videos. The proposed tool generates a keyframe-based summary, which clusters visually similar frames based on user-selected visual features and appropriate dissimilarity metrics. We discuss how this tool can be used for arthroscopic videos, taking advantage of several domain-specific aspects, without losing its ability to work on general-purpose videos. Experimental results confirm the feasibility of the proposed approach and encourage extending it to other application domains.     

Latent overlapped fingerprint separation: a review

Fingerprint matching is a widely used process to aid in crime scene investigation, where fingerprint fragments are often found on objects and surfaces. In such cases, the lifted fingerprints (called latents) are usually of poor quality and often appear overlapped and against a noisy background. These aspects make latent fingerprint – especially overlapped latent fingerprints – segmentation and enhancement (for subsequent matching) a difficult problem, for which several solutions have been proposed during the past few years. This paper presents an overview of contemporary techniques for overlapped fingerprint separation in the context of latent overlapped fingerprint matching. In addition to explaining the main concepts and surveying the literature in the field, it highlights the importance of the overlapped fingerprint segmentation (ROI extraction) process, a step for which there are no automatic techniques yet.     

Dropout vs. batch normalization: an empirical study of their impact to deep learning

Multimedia Tools and Applications - Overfitting and long training time are two fundamental challenges in multilayered neural network learning and deep learning in particular. Dropout and batch...     

New approaches to encryption and steganography for digital videos

In this work we propose a novel type of digital video encryption that has several advantages over other currently available digital video encryption schemes. We also present an extended classification of digital video encryption algorithms in order to clarify these advantages. We analyze both security and performance aspects of the proposed method, and show that the method is efficient and secure from a cryptographic point of view. Even though the method is currently feasible only for a certain class of video sequences and video codecs, the method is promising and future investigations might reveal its broader applicability. Finally, we extend our approach into a novel type of digital video steganography where it is possible to disguise a given video with another video.