In addition to revealing that the system is under development and can naturally remove specific objects in the videoProject Cloak FeaturesAt its MAX 2017 preview event, Adobe also revealed Project Scribbler, a feature that also leverages Sensei AI technology to enable natural coloring of black and white sketches using deep learning. Adobe also introduced Project Lincoln, a feature for creating visually-oriented data charts, and Project Playful Platte, which simulates real-world color palettes to further accelerate the coloring process.
Codenamed Project Scribbler, the feature currently lacks a specific service. It also utilizes Sensei artificial intelligence technology, which compares multiple images to determine the appropriate colorization results for a black and white sketch or photo. This includes rendering skin tones, hair color, and eye color in natural tones.
Furthermore, users can capture textures from specific image sections and apply them to specific locations in a black and white image, creating a natural colorization effect. For example, capturing crocodile and snakeskin textures separately can create a natural textured tone on a simple black and white bag. Combined with the previously available colorization feature for black and white sketches, creators can quickly combine new image content, reducing the time required.
While similar applications have previously appeared in other software, Adobe explains that due to its own Sensei deep learning capabilities, it's not limited to images composed of standard black and white lines, but can also be applied to real photos, with the added learning data allowing for even more natural rendering. Furthermore, Adobe emphasizes that Sensei has more data to learn from, and its learning results are accumulated almost daily, resulting in even more natural coloring results.
Adobe further explained that Project Scribbler's color decisions are pre-determined by the Sensei learning mechanism, which may use line composition to determine similarity with previously learned image content. For example, in an image similar to Einstein, it will automatically determine white skin color and silver hair color. However, users can actually make changes later on their own, and even adjust different coloring styles such as oil painting or watercolor.
Project Scribbler primarily aims to help creators accelerate their production workflows. Therefore, it's not envisioned as a replacement for existing creative processes, but rather as a suitable auxiliary tool. While Project Scribbler currently only learns specific features like faces, it will be expanded to include full-body image learning in the future, and even include content beyond portraits.
More preview features: Project Lincoln, Project Playful Plette
In other preview content, Adobe proposed Project Lincoln, a concept function to meet the design needs of creating exquisite charts. It overturns the past practice of prioritizing data and only starting to beautify the relevant charts after they are created. Instead, it allows users to complete the drawing of the chart first, and then apply the relevant data to the chart template, and make subsequent adjustments. This improves the traditional limitation of only being able to design standard charts with tools such as Excel and Novel, and solves the trouble of having to manually adjust each one when creating more exquisite charts through Illustrator CC.
Pull out the chart format, then enter the data to automatically generate the chart results. The function code-named Project Playful Plette is designed for the coloring needs of mobile devices. It allows users to select multiple sets of colors at once through a virtual palette and simulate the color mixing effect of a real palette. Users can not only quickly switch between different coloring contents, but also further use the color mixing function and even adjust the color tone of the entire painting in one go, thereby improving the trouble of traditional color changing that requires frequent clicks and multiple procedures.
As previously stated, the preview content is still in the experimental stage, so the actual corresponding service items and the specific launch time have not yet been determined.
