- The technique greatly decreases picture size, making it feasible to save on the blockchain.
- Researchers from ByteDance and the Technical University Munich created TiTok.
The new image compression method TiTok receives Ethereum co-founder Vitalik Buterin’s stamp of approval due to its promising blockchain applications. The new TiTok compression technique greatly decreases picture size, making it more feasible to save on the blockchain—but don’t get this mixed up with the social networking app TikTok.
On Farcaster, a decentralized social media platform, Buterin brought attention to TiTok’s blockchain potential by saying, “320 bits is basically a hash. Small enough to go on chain for every user.” There may be far-reaching consequences for NFTs and digital image storage of profile pitcures (PFPs) as a result of this breakthrough.
Employing Machine Learning and Sophisticated AI
Researchers from ByteDance and the Technical University Munich created TiTok, a method that compressed high-quality images into 32 bits. A 256×256 pixel picture may be reduced to “32 discrete tokens.” This is by by using sophisticated artificial intelligence (AI) image compression, as stated in the TiTok research report.
Moreover, the TiTok framework “breaks grid constraints existing in 2D tokenization methods,” allowing for more flexible and compact pictures in 1 dimension (1D). In order to tokenize photos, TiTok employs machine learning and sophisticated AI in the form of transformer-based models.
Furthermore, the technique finds and makes use of redundant information in various parts of the picture to decrease the total data size of the final result; this is called area redundancy. The research claims that TiTok’s “compact latent representation” can outperform “conventional techniques” in terms of efficiency and effectiveness.
The innovative AI-driven picture reduction approach gains confidence from the Ethereum co-founder, who emphasizes TiTok’s blockchain potential.
Highlighted Crypto News Today:
