MediaTek introduced a revolutionary text-to-image tool at MWC 2024. Here are our initial impressions.
Following the conclusion of MWC 2024 a week ago, among the plethora of intriguing products and innovations, one showcase that left a lasting impact was MediaTek’s Text-to-AI Image generation tool.
During MWC, MediaTek exhibited their SDXL Turbo AI engine, capable of converting text into images. While the concept of AI models producing images from text is not new, the innovation here lies in the tool’s offline and on-device functionality. The engine, featured in a video, runs on a prototype device powered by the Dimensity 9300 chipset without the need for an internet connection.
The AI engine’s remarkable performance enables real-time text-to-image conversion, responding in less than a second to each input prompt. The initial image generated in the demonstration was impressively detailed, depicting a cat adorned with a hat while consuming cheese. Furthermore, the engine’s agility in adapting to prompt changes stood out.
Even prominent online AI tools like ChatGPT and Google Gemini fall short of this level of efficiency. While they offer text-to-image capabilities, MediaTek’s tool, integrated into its chipset for offline use, presents a step forward in this field. Despite some limitations, such as difficulty processing prompts exceeding two characters, the engine shows great potential.
It is worth noting that this is currently a demonstration, and there is no timeline from MediaTek on its availability for smartphones equipped with the Dimensity 9300 chip. Updates will be provided as more information becomes available. Feel free to share your thoughts on MediaTek’s text-to-image tool and how you envision its utility in daily life in the comment section below!
