jakiro
Author: Josh O’Sullivan, CoinTelegraph; Compilation: White Water, Bitchain Vision
Ethereum co-founder Vitalik Buterin has endorsed the newToken for Image Tokenizer (TiTok) compression method for its potential blockchain applications.
Don’t be confused with the social media platform TikTok, the new TiTok compression method significantly reduces the image size, making it more suitable for storage on the blockchain.
Buterin highlighted TiTok’s blockchain potential on the decentralized social media platform Farcaster, saying “320 bits are basically a hash value. Small enough to get every user on the chain.”
This development could have a significant impact on the digital image storage of profile pictures (PFPs) and non-fungible tokens (NFTs).
TiTok Image Compression
TiTok was developed by ByteDance and researchers from the Technical University of Munich.The image can be compressed into 32 small data blocks (bits) without losing quality.
According to TiTok’s research paper, advanced artificial intelligence (AI) image compression allows TiTok to compress 256×256 pixel images into “32 discrete markers.”
TiTok is a one-dimensional (1D) image tokenization framework that “breaks the grid constraints present in the 2D tokenization approach” to produce a more flexible and compact image.
“So it can significantly speed up the sampling process (e.g., 410 times faster than DiT-XL/2) while achieving competitive generation quality.”
TikTok research paper shows comparisons of image compression sizes.Source: TikTok
Machine Learning Images
TiTok uses machine learning and advanced artificial intelligence to convert images into tokenized representations using transformer-based models.
This method uses region redundancy, which means it identifies and uses redundant information in different regions of the image to reduce the overall data size of the final product.
“Recent advances in generative models highlight the important role of image markers in efficient synthesis of high-resolution images.”
According to the research paper, TiTok’s “compact latent representation” can produce “more efficient and more efficient representation than traditional technology.”
Image reconstruction (a) and generation (b) using the TiTok framework (c).Source: TiTok
TikTok, not TikTok
Despite the similar names, social media platform TikTok has not been recognized by Buterin.
Ethereum co-founders highlighted TiTok’s blockchain potential, adding credibility to this new AI-powered approach to image compression.
“Unlike existing 2D VQ models that treat image latent space as 2D mesh, we provide a more compact formula to mark images as 1D latent sequences.”
The proposed new approach could “represent images with 8 to 64 times less markers than “2D markers”, and the team hopes that this study can provide inspiration for “more effective image representations.”
