<p>While technology companies are working hard to refine the speech recognition of digital voice assistants to understand the users' command and context better, Google engineers have come up with a ground-breaking Artificial Intelligence (AI)-a powered tool that can conjure images by reading just the text description.</p>.<p>It has to be noted that a private company DALL-E had first made significant development in this AI-based text-to-image generator model, but Google's latest Imagen tool has taken a huge lead in the field of photorealistic image creation. </p>.<p>Some of the sample photos shared by Google are just extraordinary. A notable aspect of the Google Imagen is that resultant photos feature real subjects such as dogs, pandas, or raccoons with real objects, not animation. </p>.<p>The kind of coherency in terms of accurately depicting exactly what is stated in the text is just marvelous. It is as close to how human imagination works in the brain while reading a book.</p>.<p>"Imagen builds on the power of large transformer language models in understanding text and hinges on the strength of diffusion models in high-fidelity image generation. Our key discovery is that generic large language models, pretrained on text-only corpora, are surprisingly effective at encoding text for image synthesis: increasing the size of the language model in Imagen boosts both sample fidelity and image-text alignment much more than increasing the size of the image diffusion model," Google Research's Brain team said.</p>.<p>This is just the start; the company says it working on improving Imagen to understand the sensitivity of the society in terms of social stereotypes, oppressive viewpoints, and derogatory, or otherwise harmful.</p>.<p>So, for now, Google has decided not to offer Imagen's programming code or a public demo, until it refines the AI algorithm to remove social bias. Once it is ready with the fully tested model, many would like Google to this feature through an application on phones and computers. Interested readers can check out the official Imagen webpage (<a href="https://imagen.research.google/" target="_blank">here</a>), to check out new sample images.</p>.<p><strong>Must read |</strong> <a href="https://www.deccanherald.com/business/technology/google-photos-app-brings-pixel-6-s-real-tone-feature-to-iphones-android-devices-1112608.html" target="_blank">Google Photos app brings Pixel 6’s Real Tone feature to iPhones, Android devices</a></p>.<p><em>Get the latest news on new launches, gadget reviews, apps, cybersecurity, and more on personal technology only on <a href="https://www.deccanherald.com/tag/dh-tech?_ga=2.210580691.73733284.1595225125-1706599323.1592232366" target="_blank">DH Tech.</a></em> </p>
<p>While technology companies are working hard to refine the speech recognition of digital voice assistants to understand the users' command and context better, Google engineers have come up with a ground-breaking Artificial Intelligence (AI)-a powered tool that can conjure images by reading just the text description.</p>.<p>It has to be noted that a private company DALL-E had first made significant development in this AI-based text-to-image generator model, but Google's latest Imagen tool has taken a huge lead in the field of photorealistic image creation. </p>.<p>Some of the sample photos shared by Google are just extraordinary. A notable aspect of the Google Imagen is that resultant photos feature real subjects such as dogs, pandas, or raccoons with real objects, not animation. </p>.<p>The kind of coherency in terms of accurately depicting exactly what is stated in the text is just marvelous. It is as close to how human imagination works in the brain while reading a book.</p>.<p>"Imagen builds on the power of large transformer language models in understanding text and hinges on the strength of diffusion models in high-fidelity image generation. Our key discovery is that generic large language models, pretrained on text-only corpora, are surprisingly effective at encoding text for image synthesis: increasing the size of the language model in Imagen boosts both sample fidelity and image-text alignment much more than increasing the size of the image diffusion model," Google Research's Brain team said.</p>.<p>This is just the start; the company says it working on improving Imagen to understand the sensitivity of the society in terms of social stereotypes, oppressive viewpoints, and derogatory, or otherwise harmful.</p>.<p>So, for now, Google has decided not to offer Imagen's programming code or a public demo, until it refines the AI algorithm to remove social bias. Once it is ready with the fully tested model, many would like Google to this feature through an application on phones and computers. Interested readers can check out the official Imagen webpage (<a href="https://imagen.research.google/" target="_blank">here</a>), to check out new sample images.</p>.<p><strong>Must read |</strong> <a href="https://www.deccanherald.com/business/technology/google-photos-app-brings-pixel-6-s-real-tone-feature-to-iphones-android-devices-1112608.html" target="_blank">Google Photos app brings Pixel 6’s Real Tone feature to iPhones, Android devices</a></p>.<p><em>Get the latest news on new launches, gadget reviews, apps, cybersecurity, and more on personal technology only on <a href="https://www.deccanherald.com/tag/dh-tech?_ga=2.210580691.73733284.1595225125-1706599323.1592232366" target="_blank">DH Tech.</a></em> </p>