Search Results (2)

Biologically inspired retinal preprocessing improves visual perception by efficiently encoding and reducing entropy in images. In this study, we introduce a new saliency prediction framework that combines a retinal model with deep neural networks (DNNs) using information theory ideas. By mimicking the human retina, our method creates clearer saliency maps with lower entropy and supports efficient computation with DNNs by optimizing information flow and reducing redundancy. We treat saliency prediction as an information maximization problem, where important regions have high information and low local entropy. Tests on several benchmark datasets show that adding the retinal model boosts the performance of various bottom-up saliency prediction methods by better managing information and reducing uncertainty. We use metrics like mutual information and entropy to measure improvements in accuracy and efficiency. Our framework outperforms state-of-the-art models, producing saliency maps that closely match where people actually look. By combining neurobiological insights with information theory—using measures like Kullback–Leibler divergence and information gain—our method not only improves prediction accuracy but also offers a clear, quantitative understanding of saliency. This approach shows promise for future research that brings together neuroscience, entropy, and deep learning to enhance visual saliency prediction. Full article

Chromatin remodeling complexes are required for many distinct nuclear processes such as transcription, DNA replication, and DNA repair. However, the contribution of these complexes to the development of complex tissues within an organism is poorly characterized. Imitation switch (ISWI) proteins are among the most evolutionarily conserved ATP-dependent chromatin remodeling factors and are represented by yeast Isw1/Isw2, and their vertebrate counterparts Snf2h (Smarca5) and Snf2l (Smarca1). In this study, we focused on the role of the Snf2h gene during the development of the mammalian retina. We show that Snf2h is expressed in both retinal progenitors and post-mitotic retinal cells. Using Snf2h conditional knockout mice (Snf2h cKO), we found that when Snf2h is deleted, the laminar structure of the adult retina is not retained, the overall thickness of the retina is significantly reduced compared with controls, and the outer nuclear layer (ONL) is completely missing. The depletion of Snf2h did not influence the ability of retinal progenitors to generate all the differentiated retinal cell types. Instead, the Snf2h function is critical for the proliferation of retinal progenitor cells. Cells lacking Snf2h have a defective S-phase, leading to the entire cell division process impairments. Although all retinal cell types appear to be specified in the absence of the Snf2h function, cell-cycle defects and concomitantly increased apoptosis in Snf2h cKO result in abnormal retina lamination, complete destruction of the photoreceptor layer, and consequently, a physiologically non-functional retina. Full article