aiTransformer

      aiTransformer added a new enhance type - AI enhance with face restoration, which is based on an AI algorithm particularly good with faces. It can help you make flawless selfies.     Up to now, aiTransformer has 5 different enhance types: AI overall enhance,  detail enhance, sharpen, brighten, and AI enhance with face restoration. Every type has its own strength and weakness.      AI overall enhance, brighten are general enhancers, they can be applied to all images: portraits, landscapes, buildings... Results are generally decent.      Detail Enhance magnifies details, Sharpen  improves edges.  This makes them good for enhancing landscapes, objects, etc...But if they are applied on faces, the results won't be pretty. Comparison of the original, detail enhance and sharpen on faces   Comparison of the original, detail enhance and sharpen on objects       AI enhance with face restoration does overall enhancement and in particular restores faces when present.       Here are some e

      aiTransformer has a new tool: the Super Stylizer, which can generate stunning detailed images conditioned on text descriptions and use the generated images to stylize your picture with adjustable style strength. It is based on Stable Diffusion. How to use the Super Stylizer What is Stable Diffusion     Stable Diffusion is probably the hottest AI algorithm recently, it is an open-source, deep learning, text-to-image model, which can generate images from text or modify images based on text.      Released in 2022, Stable diffusion is developed by Compvis (the Computer Vision research group at Ludwig Maximilian University of Munich) and sponsored mainly by Stability AI. It has been trained on billions of images and can produce impressive results. What is a prompt           In order to produce desired results, it's important to write a good prompt. A prompt is the description of the image the AI will generate.   How to write a Stable Diffusion prompt       With the help of AI, an

We added three more new cartoon styles: Celeba Distill, Face Paint 1, Face paint 2.  These styles work very well on "big head" photos with only one face. The new styles also work fine for big head photos with more than one face. But, here comes the BUT part, if the faces are not big enough, the results can be distorted. What will happen to pictures without any faces? It's fun to try.                                               The same rules apply to the Video Cartoonizer. They work well with big head videos. Beware the new styles take much longer to process videos.                                             Celeba Distill                                            Face Paint 1                                             Face Paint 2 If the face is not big enough, transformed videos can be distorted.                                                                 Face paint 1 and pace paint 2 have similar issues, probably because the AI those new styles are based on w

       Have you noticed we have a new sticky menu and new batch options for filters in MultiStylizer? Those are not all, we also added four new Japanese Cartoon styles to both Cartoonizer and Video Cartoonizer.      Now, there are five cartoon styles you can choose from. After selecting a style, an example will show how it looks like. Examples for cartoonized videos                                                                            White-box                                                                                                                                             Hayao                                                                          Hosoda         Paprika                                                                             Shinkai

  MultiStylizer: The Joy of Life by Matisse, Filter: Cool aiTransformer added a new feature in  MultiStylizer: Filters. It provides 16 different filters.                                        If no filter is selected, the result will be transformed as before. MultiStylizer: The Joy of Life by Matisse When a filter is picked,  an example of the filter effect will show and the filter will be applied to the result during processing. The filters can also be applied independently to the original picture, without any style images selected.  MultiStylizer Filter: Sepia For comparison,   Sidelights: A story about how the filters were added.    Our developer accidentally changed the colour codes when he was trying to fix a bug. A fascinating result was showed later in testing.  No one knew what happened at first, AI evolved? It was just kidding. However, it did take us sometime to finally figure out it was only a new bug being produced when fixing an old one. Phew~~~ Since the bug produced a

    aiTransformer added another new function: Video Enhancer , which applies the enhancer function frame by frame and makes them sharper and more vivid, with the original audio intact.       The original video for comparison.                                           

  Added an animated gif output in Stylizer to show the style transitions from low to high intensity. Added a special message feature, also enhanced error handling and information presentation.

           aiTransformer Stylizer produces 6 stylized results with various style intensity applied, now a new result is added: a GIF image showing the style transition from low to high intensity.        It shows as the first result.         Now you can make animated stylized pictures. Try it out, it's fun                                     

         Cartoon Version of Charade by Cary Grant and Audrey Hepburn 1963 We just added a new function in AI Image Transformer :  Video Cartoonizer , which I'm quite excited about.  Try it out, make your own cartoon movies. It's easy to use, select the video you want to turn to a cartoon movie, "upload to process". It could take a few minutes to transform, depends on the size and the number of frames in the video. The limit of the video size is 100M. If you have a larger video, split it into several smaller ones first. For long running processes, the progress bar may not be updated correctly due to interrupted communication link between client (your browser or app) and the server, probably because of internet fluctuation. Then the result fails to show. It's rare, but it happens. In this case, you need to redo the "Upload to Process". In most cases the server has already done the work and the result is already there, then the redo will be fast. A tip fo

- Added Video Cartoonizer function that can make cartoon movies directly from home videos. - Enhanced the backend with GPU support, now some functions can run many times faster than before, making the Video Cartoonizer practical. - Improved error handling, especially for large image files that can cause out of memory errors.

    aiTransformer just added a new feature: Apply Styles to Regions, in both Stylizer and Multistylizer.        There are three options:"All", "Foreground" and "Background". “All" is the default option, it applies the chosen style to the whole image.   Multistylizer:Bamboo by Shi Tao, All     "Foreground" only applies the chosen style to the forepart and keeps the background unchanged. Multistylizer: Woodcutter by Kazimir Malevich, Foreground Stylizer:The Great Wave off Kanagawa by Hokusai, Foreground                                         "Background" does the apposite, applies the chosen style to the background and keeps the foreground unchanged.   Stylizer:The Birth of Venus by Sandro Botticelli, Background     Let's try something fun, apply one style image to the foreground and another style image to the background.     This image was transformed twice:            Step 1. apply the style image Music by Dorrit Black to for

 Enhanced Stylizer and MultiStylizer with the option to apply styles to the whole region, or the auto-detected foreground / background region, also added more predefined styles and batch options.

Multistylizer: Long Grass with Butterflies by Van Gogh     Multistylizer: Stained Glass Mosaic of A Lady Holding A Flower

    aiTransformer has two ways to transfer images to paintings: Stylizer and Multistylizer .  Stylizer can only take one style image, while Multistylizer can take up to three style images.      What will happen if we only use one style image in Multistylizer? Will it produce the same result as the Stylizer does? The answer is "No".  Stylizer and Multistylizer use different AI algorithms, therefore, even  using the same style image to transfer the same photo, they will produce different results.          Sometimes, the results are significantly different.

Redesigned the backend so that the functions can run faster and with consistent results. Added Gallery function to showcase creations using the app, with the newest image showing on the homepage as Transformation of The Day. New images will be added to the gallery regularly. Added a Chatbot and link to the transformer blog.  

“Everyone is an artist."                                    ------------Joseph Beuys     We all have an artist inside, but not every one has the skills to express his/her innate creativity as art works. Fortunately, using different algorithms,  AI can help us to turn photos into art pieces.     aiTransformer is an all-in-one, easily operating image transformation app,  using AI algorithms to enhance, enlarge and transform any images into cartoons, sketches, and paintings       aiTransformer has six functions: Cartoonizer , Enhancer , Sketcher , Stylizer , Multistylizer , and Enlarger . You can either click the blue button in each block to go function pages or use the menu. How to Transfer Photos to Cartoons         Want to use your own cartoon avatars? Cartoonizer is where you go.      "Choose files" to pick the photo you want to transfer, then click "upload to process".          Download the result images if you like them. The images are only avail