Deep Salience: Visual Salience Modeling via Deep Belief Propagation

Visual salience is an intriguing phenomenon observed in biological neural systems. Numerous attempts have been made to model visual salience mathematically using various feature contrasts, either locally or globally. However, these algorithmic models tend to ignore the problem’s biological solutions, in which visual salience appears to arise during the propagation of visual stimuli along the visual cortex. In this paper, inspired by the conjecture that salience arises from deep propagation along the visual cortex, we present a Deep Salience model where a multi-layer model based on successive Markov random fields (sMRF) is proposed to analyze the input image successively through its deep belief propagation. As a result, the foreground object can be automatically separated from the background in a fully unsupervised way. Experimental evaluation on the benchmark dataset validated that our Deep Salience model can consistently outperform eleven state-of-the-art salience models, yielding the higher rates in the precision-recall tests and attaining the best F-measure and mean-square error in the experiments.

Outcomes of surgical bleb revision for complications of trabeculectomy.

OBJECTIVE: To describe the results of revision surgery for complications of trabeculectomy in a case series from an academic glaucoma service. DESIGN: Retrospective case series. PARTICIPANTS: A total of 177 eyes of 167 adult patients who underwent revision of trabeculectomy at the Wilmer Eye Institute between 1994 and 2007. METHODS: Three indications for surgery were identified: hypotony without leak, bleb leak, and bleb dysesthesia. Revision was deemed successful when all of the following were true: the primary indication was eliminated, further intraocular pressure (IOP)-lowering surgery was not required, no major complication occurred, and a new bleb-related problem did not develop. Patients with less than 3 months of follow-up were excluded unless failure occurred earlier. Surgical procedures included variations on excision of thin or leaking conjunctiva with advancement. MAIN OUTCOME MEASURES: Change in IOP, change in visual acuity, need for further IOP-lowering surgery, and complications after bleb revision. RESULTS: Subjects' mean age was 67+/-14 years, 54% were female, and mean follow-up was 2.8+/-2.7 years, with a mean interval from trabeculectomy to revision of 3.5+/-3.7 years. Overall success rate was 63% (112/177), which was slightly higher for leak repair (65%; 64/98) and hypotony (63%; 32/51) than for dysesthesia (57%; 16/28) indications. By Kaplan-Meier analysis, overall cumulative success rates at 1, 2, 5, and 10 years after bleb revision were 80%, 75%, 50%, and 41%, respectively. IOP and visual acuity improved significantly in both hypotony and leak groups (P values ranging from 0.004 to <0.0001). Additional IOP-lowering surgery was required in 9%. In multivariate regression analysis adjusting for age, gender, and number of prior surgeries, patients with glaucoma other than primary open-angle glaucoma were twice as likely to have failed bleb revision. CONCLUSIONS: Surgical bleb revision often provides successful resolution of bleb-related complications. Most patients maintain IOP control without need for further IOP-lowering surgery. FINANCIAL DISCLOSURE(S): Proprietary or commercial disclosure may be found after the references.