STEM

Fun with AI: Text and Image->Video

One of the relatively newer tools that came online for the general public designed to transform ideas into AI-generated videos is Hailuo AI. In this post, I’ll share a couple of examples of its outputs based on my paintings, drawings, digital creations, and text prompts.

Hailuo’s Conch Video is a product of Minimax, a China-based AI company. It offers a range of features for video creation, including:
Text to Video: Convert text prompts into engaging video content.
Image to Video: Transform images into dynamic video sequences.
Customizable Templates: Use user-friendly templates to create unique videos.
Automated Video Editing: Includes tools for video trimming, intelligent transitions, and deep learning video generation.

Very quickly, Hailuo AI has gained immense popularity among content creators, marketers, e-learning platforms, and hobbyists. Currently, it’s offering a 3-day free trial that allowes creation of 6-second clips based on text prompts and static images along with text prompts. To create longer videos, you must generate multiple clips and stitch them together using another editing tool. Also, you’ll need to use a different tool for adding audio to video and generate the combined video outputs. More info @ https://hailuoaivideo.net/

Let me share some example video clips I generated. The following one was created by text prompts, in two separate sessions. Then using another video editing software, they were combined, and I added some sound effects and text to the combined video. It’s a message of peace.

Along this line, if you’re interested in listening to a song I created titled “Peace for Humanity“, you can check it out here (it has the full song 3:27s with full lyrics). To just hear the audio, use the player below. The vocal is also AI-generated. Amazing!


The following one is also generated by prompts where the premise is a puppy looking for its owner in a public city park. I didn’t add any audio to this, but the rendering and interpretation of the prompt was impressive as you can see below.


The next examples are based on static image uploads with no audio. For example, one of my canvas paintings (a young womanl holding a peach) was scanned and uploaded to the AI platform, then I asked it to create a video of the painting subject (the woman) eating the peach and smile. One version, despite the same prompts, was more realistic to the painting, while the other one is more cartoonish as you can see below.


Here’s another clip generated from a static image. First, the image was rendered in DallE, then it was fed to Hailuo AI with directions to make the “alien girl” be angry and move about specifically. Try not to get startled 😉


Below is a Shoebill closeup looking around.


Here’s a made-up character digitally drawn by me that looks like a marshmallow with tentacles, and it’s animated using prompts. It’s made to smile, talk and wiggle its tentacles. No audio added.

The following video was generated based on my painting of Sandpipers on canvas (scanned to digital) and then with text prompts, I animated them using AI to generate the video, then manually edited the video to embed audio to generate the final short clip. The first frame shows the oiriginal painting, then the following frames are AI-generated animation frames.

Overall, I was very impressed with the model’s prompt interpretation, the rendering AI’s ability to tackle lighting, motion, and most importantly derive a 3D object’s proportions, depths etc. of even ones it had never seen before during training. For example, a portrait of a human, animal, and common objects can be more easily deciphered from an image and rendered for 3D animation, but creatures, like some my fantasy characters that I used during my experiments were really impressive…in fact, this is the best AI model for rendering image->video, and text->video because of its superior abilities that’s out there today. I expect Hailuo is only going to get more polished, unless bought off and killed by a company or regulations, as they incorporate audio, AI transcripting, AI story-writing, AI voice to text to video, and advanced video editing capabilities.

I hope it was a fun read for you. Happy creating!

Back To Top