2023年对于AI视频领域来说，被认为是一个具有里程碑意义的一年。在短短12个月的时间内，AI视频技术取得了令人瞩目的突破。年初时，公开的文本转视频模型尚不存在，而现在，数十种视频生成产品已经被广泛应用，全球数百万用户通过简单的文本或图像提示，就能轻松创建出短片。
然而，尽管取得了巨大的进步，这些产品仍然存在一些局限性。目前，大多数生成的视频时长仍然只有3到4秒，而且输出的质量也参差不齐。例如，角色的一致性等问题还未得到有效解决。这也意味着，我们距离仅通过单一文本提示，就能创造出皮克斯级别的短片，还有很长的路要走。
This is a breakthrough year for AI video in 2023. Within just 12 months, AI video technology has made remarkable progress. At the beginning of the year, there were no publicly available text-to-video models, but now, dozens of video generation products are widely used, with millions of users around the world creating short videos through simple text or image prompts.
However, despite these significant advancements, there are still limitations with these products. Currently, most generated videos are still only 3 to 4 seconds long, and the quality of the outputs varies. Issues such as character consistency have not yet been effectively resolved. This also means that we still have a long way to go before we can create Pixar-level shorts solely through a single text prompt.
【来源】https://mp.weixin.qq.com/s/_WdPoPfptK8zch7C4dh1iQ
Views: 1