Stability AI Launches Comprehensive API Service, Leading a New Era in Image Processing

Hong Kong and UK Research Teams Propose Innovative Image Tokenization Method; Hierarchical Structure Improves Reconstruction Quality

Researchers from Hong Kong and the UK have recently proposed a novel image tokenization method designed to convert images into digital representations (tokens) in a more compact and accurate manner. Unlike traditional methods that distribute information uniformly across all tokens, this method employs a hierarchical structure, capturing visual information layer by layer, thereby improving the quality and efficiency of image reconstruction. Traditional image tokenization techniques typically divide each part of an image equally into multiple tokens, whereas the new method adopts a hierarchical structure. Initial tokens encode rough shapes and structural elements,

Li Fei-Fei's Team Unveils Novel Image Processing Technology, Breaking Traditional Boundaries

Stability AI Releases Stable Virtual Camera: Effortlessly Transform 2D Photos into 3D Videos

AI Digital Human Breakthrough! Hedra Launches Character-3 Model and Hedra Studio: Joint Reasoning on Images, Text, and Audio

Hedra has announced a significant advancement in AI digital humans with the launch of its Character-3 model and Hedra Studio. This new technology enables joint reasoning across images, text, and audio, leading to more realistic and interactive digital human experiences.

Stability AI Partners with Arm to Deliver Offline AI Audio Generation

Microsoft Paint App Introduces Free AI Eraser Feature for Easy Element Removal!

Microsoft has recently upgraded its classic Paint app with a new AI-powered eraser feature, making it easier for users to work with images. The core of this feature lies in the ability for users to simply circle the elements they want to delete, and the AI will automatically recognize and erase them, significantly simplifying the traditional image processing workflow. After two months of testing, this new feature has been officially rolled out to all users. Users can enjoy this feature for free by upgrading to the latest version in the Microsoft Store. However, it is worth noting that because the AI model requires

Stability AI Launches SPAR3D: Generate 3D Objects from a Single Image in One Second

Downtime Again? ChatGPT and Sora Experienced Outage for the Second Time This Month

On Thursday local time in the U.S., ChatGPT, Sora, and OpenAI's API services for developers encountered another outage, lasting over four hours. OpenAI reported a significant disruption starting at 11 AM (all times referenced in this article are local to the U.S.), but by 3:16 PM, services were gradually being restored. Regular users of ChatGPT might recall that there was also an outage earlier this month. On its status page, OpenAI indicated that this major disruption was caused by issues with its upstream provider.

Tsinghua University Collaborates with Tencent to Present ColorFlow: Automatically Coloring Black and White Comics While Maintaining Character Consistency

With the development of digital art, automated image processing technology has received increasing attention. Recently, a research team from Tsinghua University and Tencent's ARC Lab proposed a novel image sequence coloring model called ColorFlow. This model aims to address the challenge of maintaining consistency in character and object identity while coloring black and white image sequences, fulfilling the practical needs of industries such as comics and animation. ColorFlow is a three-stage diffusion base framework that leverages contextual information by utilizing a reference image pool for black and white.

Stability AI Launches New Model Stable Zero123: Enhancing Quality of 3D Object Generation from a Single Image

Stability AI has released a new AI model called Stable Zero123, which can generate high-quality 3D object views from a single image. Stable Zero123 significantly outperforms its predecessor, Zero123-XL, thanks to three key innovations. It uses a rigorously curated training dataset sourced from Objaverse, specifically retaining high-quality images.