Tree-based iterated local search for Markov random fields with applications in image analysis

The maximum a posteriori assignment for general structure Markov random fields is computationally intractable. In this paper, we exploit tree-based methods to efficiently address this problem. Our novel method, named Tree-based Iterated Local Search (T-ILS), takes advantage of the tractability of tree-structures embedded within MRFs to derive strong local search in an ILS framework. The method efficiently explores exponentially large neighborhoods using a limited memory without any requirement on the cost functions. We evaluate the T-ILS on a simulated Ising model and two real-world vision problems: stereo matching and image denoising. Experimental results demonstrate that our methods are competitive against state-of-the-art rivals with significant computational gain.

Our Sherlockian eyes: the surveillance of vision

or this inter-disciplinary article, we undertook a pilot case study that eye-tracked the ‘Holmes Saves Mrs. Hudson’ sequence from the episode, A Scandal in Belgravia (Sherlock, BBC, 2012). This small-scale empirical study involved a total of 13 participants (3 males and 10 females, mean age was: 27 years), comprised of a mixture of academics and undergraduate students at La Trobe University in Melbourne, Australia. The article examines its findings through a range of threaded frames – neuroscience, forensics, surveillance, haptics, memory, performance-movement, and relationality – and uniquely draws upon the interests of the authors to set the examination in context. The article is both a reading of Sherlock and a dialogue between its authors. We discover that the codes and conventions of Sherlock have a direct impact on where viewers look but we also discover eyes emerging in the periphery of the frame, and we account for these ways of seeing in different ways.

Visual impairment in Australia: distance visual acuity, near vision, and visual field findings of the Melbourne Visual Impairment Project

PURPOSE: To describe the age-specific and gender-specific rates of blindness and visual impairment in urban adults aged 40 years and older. METHODS: A population-based sample of residents was recruited. Presenting and best-corrected distance visual acuities were assessed. Functional near vision was measured at each participant's preferred distance. Visual field examination was performed with a Humphrey Field Analyzer (HFA); those unable to perform the field analyzer test attempted a Bjerrum screen or confrontation field. RESULTS: The study population comprised 3,271 residents (83% of eligible) from ages 40 to 98 years; 54% were women. Overall, 56% of the study population wore distance correction; this was significantly lower in men but higher in the older age groups. Age-adjusted rates of blindness were 0.066% in men and 0.170% in women. Vision with current correction improved after refraction by gender and age. Direct age-standardized rates of functional near vision did not vary significantly by gender. Forty-six people had significant visual field loss in their better eye. The proportion of participants with constriction of the visual field to within 20 degrees of fixation was similar for men and women when controlled for age (odds ratio, 0.81; 95% confidence interval, 0.44 to 1.49) but increased significantly with age controlled for gender. Visual field abnormalities were detected in 548 right eyes (17%) and 533 left eyes (16%). CONCLUSIONS: Although overall rates of blindness because of visual acuity loss were relatively low, nearly three times more people had visual impairment because of visual field loss than visual acuity loss. These results highlight the need to target blindness prevention programs to the aging population, with a special emphasis on women.