Arbitrary Font Generation by Encoder Learning of Disentangled Features.

Making a new font requires graphical designs for all base characters, and this designing process consumes lots of time and human resources. Especially for languages including a large number of combinations of consonants and vowels, it is a heavy burden to design all such combinations independently. Automatic font generation methods have been proposed to reduce this labor-intensive design problem. Most of the methods are GAN-based approaches, and they are limited to generate the trained fonts. In some previous methods, they used two encoders, one for content, the other for style, but their disentanglement of content and style is not sufficiently effective in generating arbitrary fonts. Arbitrary font generation is a challenging task because learning text and font design separately from given font images is very difficult, where the font images have both text content and font style in each image. In this paper, we propose a new automatic font generation method to solve this disentanglement problem. First, we use two stacked inputs, i.e., images with the same text but different font style as content input and images with the same font style but different text as style input. Second, we propose new consistency losses that force any combination of encoded features of the stacked inputs to have the same values. In our experiments, we proved that our method can extract consistent features of text contents and font styles by separating content and style encoders and this works well for generating unseen font design from a small number of reference font images that are human-designed. Comparing to the previous methods, the font designs generated with our method showed better quality both qualitatively and quantitatively than those with the previous methods for Korean, Chinese, and English characters. e.g., 17.84 lower FID in unseen font compared to other methods.

Part Affinity Fields and CoordConv for Detecting Landmarks of Lumbar Vertebrae and Sacrum in X-ray Images.

With the prevalence of degenerative diseases due to the increase in the aging population, we have encountered many spine-related disorders. Since the spine is a crucial part of the body, fast and accurate diagnosis is critically important. Generally, clinicians use X-ray images to diagnose the spine, but X-ray images are commonly occluded by the shadows of some bones, making it hard to identify the whole spine. Therefore, recently, various deep-learning-based spinal X-ray image analysis approaches have been proposed to help diagnose the spine. However, these approaches did not consider the characteristics of frequent occlusion in the X-ray image and the properties of the vertebra shape. Therefore, based on the X-ray image properties and vertebra shape, we present a novel landmark detection network specialized in lumbar X-ray images. The proposed network consists of two stages: The first step detects the centers of the lumbar vertebrae and the upper end plate of the first sacral vertebra (S1), and the second step detects the four corner points of each lumbar vertebra and two corner points of S1 from the image obtained in the first step. We used random spine cutout augmentation in the first step to robustify the network against the commonly obscured X-ray images. Furthermore, in the second step, we used CoordConv to make the network recognize the location distribution of landmarks and part affinity fields to understand the morphological features of the vertebrae, resulting in more accurate landmark detection. The proposed network was evaluated using 304 X-ray images, and it achieved 98.02% accuracy in center detection and 8.34% relative distance error in corner detection. This indicates that our network can detect spinal landmarks reliably enough to support radiologists in analyzing the lumbar X-ray images.

A Review of Multi-Modal Learning from the Text-Guided Visual Processing Viewpoint.

For decades, co-relating different data domains to attain the maximum potential of machines has driven research, especially in neural networks. Similarly, text and visual data (images and videos) are two distinct data domains with extensive research in the past. Recently, using natural language to process 2D or 3D images and videos with the immense power of neural nets has witnessed a promising future. Despite the diverse range of remarkable work in this field, notably in the past few years, rapid improvements have also solved future challenges for researchers. Moreover, the connection between these two domains is mainly subjected to GAN, thus limiting the horizons of this field. This review analyzes Text-to-Image (T2I) synthesis as a broader picture, Text-guided Visual-output (T2Vo), with the primary goal being to highlight the gaps by proposing a more comprehensive taxonomy. We broadly categorize text-guided visual output into three main divisions and meaningful subdivisions by critically examining an extensive body of literature from top-tier computer vision venues and closely related fields, such as machine learning and human-computer interaction, aiming at state-of-the-art models with a comparative analysis. This study successively follows previous surveys on T2I, adding value by analogously evaluating the diverse range of existing methods, including different generative models, several types of visual output, critical examination of various approaches, and highlighting the shortcomings, suggesting the future direction of research.

Improving the optical performance of InGaN light-emitting diodes by altering light reflection and refraction with triangular air prism arrays.

The effect of triangular air prism (TAP) arrays with different distance-to-width (d/w) ratios on the enhancement of light extraction efficiency (LEE) of InGaN light-emitting diodes (LEDs) is investigated. The TAP arrays embedded at the sapphire/GaN interface act as light reflectors and refractors, and thereby improve the light output power due to the redirection of light into escape cones on both the front and back sides of the LED. Enhancement in radiometric power as high as 117% and far-field angle as low as 129° are realized with a compact arrangement of TAP arrays compared with that of a conventional LED made without TAP arrays under an injection current of 20 mA.

Induction of neuronal vascular endothelial growth factor expression by cAMP in the dentate gyrus of the hippocampus is required for antidepressant-like behaviors.

The cAMP cascade and vascular endothelial growth factor (VEGF) are critical modulators of depression. Here we have tested whether the antidepressive effect of the cAMP cascade is mediated by VEGF in the adult hippocampus. We used a conditional genetic system in which the Aplysia octopamine receptor (Ap oa(1)), a G(s)-coupled receptor, is transgenically expressed in the forebrain neurons of mice. Chronic activation of the heterologous Ap oa(1) by its natural ligand evoked antidepressant-like behaviors, accompanied by enhanced phosphorylation of cAMP response element-binding protein and transcription of VEGF in hippocampal dentate gyrus (DG) neurons. Selective knockdown of VEGF in these cells during the period of cAMP elevation inhibited the antidepressant-like behaviors. These findings reveal a molecular interaction between the cAMP cascade and VEGF expression, and the pronounced behavioral consequences of this interaction shed light on the mechanism underlying neuronal VEGF functions in antidepression.

Enhanced air-cavity effect of periodically oriented embedded air protrusions for high-efficiency InGaN/GaN light-emitting diodes.

We report on the development of periodically oriented embedded air protrusion (EAP) structures at the GaN-sapphire interface in InGaN/GaN LEDs. A specific SiO(2) mask pattern and a simple wet etching process were utilized for the fabrication of EAP structures. A strong coupling between closely proximate air cavities and the multiple quantum wells promoted spontaneous emission due to the high-index contrast at the GaN-air interface. As a result, the light output power of the EAP LED was 2.2 times higher than that of a conventional LED at an injection current of 20 mA.

A novel conditional genetic system reveals that increasing neuronal cAMP enhances memory and retrieval.

Consistent evidence from pharmacological and genetic studies shows that cAMP is a critical modulator of synaptic plasticity and memory formation. However, the potential of the cAMP signaling pathway as a target for memory enhancement remains unclear because of contradictory findings from pharmacological and genetic approaches. To address these issues, we have developed a novel conditional genetic system in mice based on the heterologous expression of an Aplysia octopamine receptor, a G-protein-coupled receptor whose activation by its natural ligand octopamine leads to rapid and transient increases in cAMP. We find that activation of this receptor transgenically expressed in mouse forebrain neurons induces a rapid elevation of hippocampal cAMP levels, facilitates hippocampus synaptic plasticity, and enhances the consolidation and retrieval of fear memory. Our findings clearly demonstrate that acute increases in cAMP levels selectively in neurons facilitate synaptic plasticity and memory, and illustrate the potential of this heterologous system to study cAMP-mediated processes in mammalian systems.

Valproic acid promotes neuronal differentiation by induction of proneural factors in association with H4 acetylation.

Valproate (VPA) influences the proliferation and differentiation of neuronal cells. However, little is known about the downstream events, such as alterations in gene transcription, that are associated with cell fate choice. To determine whether VPA plays an instructive role in cell fate choice during hippocampal neurogenesis, the expression of genes involved in the cell cycle and neuronal differentiation was investigated. Treatment with VPA during the progenitor stages resulted in strong inhibition of cell proliferation and induction of neuronal differentiation, accompanied by increases in the expression of proneural transcription factors and in neuronal cell numbers. The increased expression of Ngn1, Math1 and p15 points to a shift towards neuronal fate in response to histone deacetylase inhibitors (HDACi). Chromatin immunoprecipitation (ChIP) analysis showed that acetylated histone H4 (Ac-H4) was associated with the Ngn1, Math1 and p15 promoters in cultured hippocampal neural progenitor cells. VPA-induced hippocampal neurogenesis was also accompanied by association of Ac-H4 with the Ngn1 promoter in hippocampal extracts. The discovery of an association between HDACi and the Ngn1, Math1 and p15 promoters extends the importance of HDAC inhibition as a key regulator of neuronal differentiation at the transcriptional level.

Activation of NMDA receptors increases proliferation and differentiation of hippocampal neural progenitor cells.

The prolonged effects of N-methyl-D-aspartate (NMDA) receptor activation on the proliferation and differentiation of hippocampal neural progenitor cells (NPCs) were studied. Under conditions of mitogen-mediated proliferation, a single NMDA pulse (5 microM) increased the fraction of 5-bromo-2-deoxyuridine (BrdU)-positive (BrdU(+)) cells after a delay of 72 hours. Similarly, a single systemic injection of NMDA (100 mg/kg) increased the number of BrdU(+) cells in the dentate gyrus (DG) after 28 days, but not after 3 days. NMDA receptor activation induced an immediate influx of Ca(2+) into the NPCs and the NPCs expressed and released vascular endothelial growth factor (VEGF) in an NMDA receptor-dependent manner within 72 hours. With repetitive stimulation at the same dose, NMDA stimulated the acquisition of a neuronal phenotype accompanied by an increase in the expression of proneural basic helix-loop-helix (bHLH) factors. Together these findings suggest that neurogenesis in the developing brain is likely to be both directly and indirectly regulated by complex interactions between Ca(2+) influx and excitation-releasable cytokines, even at mild levels of excitation. In addition, our results are the first to show that stimulation of NPCs may lead to either proliferation or neuronal differentiation, depending on the level of NMDA receptor activation.