How to Create a Realistic 10-Second AI Cat and Mouse Trap Video From One Image
Artificial intelligence has made it possible to transform a single still image into an engaging video with realistic movement, cinematic camera work, lighting, and sound. One creative example is a scene showing an orange tabby cat watching a small mouse near a woven basket and a piece of cheese.
The idea is to take the original image and animate it without changing the important elements. The cat should remain the same, the mouse should remain the same, the basket should stay in its original position, and the kitchen environment should remain consistent. Only subtle movement is introduced to make the photograph feel alive.
Modern AI video models can generate videos from an input image combined with a text prompt. Google's current Veo documentation, for example, describes image-to-video generation, reference images, camera controls, realistic physics, and improved prompt adherence. �
Google DeepMind
The Concept Behind the Video
The scene begins with an orange tabby cat carefully watching a small mouse near a woven basket. A piece of cheese is placed on the floor nearby.
Rather than creating a violent scene, the video focuses on a playful cat-and-mouse moment. The cat carefully moves around the basket while watching the mouse, creating the impression that it is trying to prevent the mouse from escaping.
The mouse notices the cat and moves cautiously toward the cheese. The cat follows its movement with its eyes and makes small natural movements with its paws, ears, and whiskers.
Because the video is only 10 seconds long, the action should remain simple. A short and clear sequence generally gives an AI video generator fewer complicated movements to interpret.
Complete 10-Second AI Video Prompt
Use the following prompt with an image-to-video generator that accepts a reference image:
Create a highly realistic 10-second video using the uploaded image as the exact reference. Do not change anything in the original image: keep the exact same orange tabby cat, same face, fur, body, basket, mouse, cheese, kitchen, lighting, colors, and composition. The cat remains focused on the mouse and carefully moves around the basket as if setting a clever trap. The mouse slowly moves toward the cheese and then pauses after noticing the cat. Keep every character and object exactly the same throughout the entire video. Only add natural, realistic movement: subtle cat breathing, blinking, ear movement, whisker movement, gentle paw movement, and small natural mouse movement. Use a very slow cinematic camera push-in while preserving the original scene. Ultra-realistic fur, natural physics, realistic shadows and reflections, smooth professional cinematic quality, 4K, vertical 9:16. No new objects, no removed objects, no face changes, no body changes, no color changes, no distortion, no text, and no watermark.
This prompt is designed around several important elements: reference-image consistency, character movement, camera movement, lighting, realism, and restrictions on unwanted changes.
Why the Original Image Should Be Used as a Reference
The reference image is extremely important when the goal is to preserve the appearance of the cat, mouse, basket, cheese, and kitchen.
Instead of asking the AI to recreate the entire scene from a text description, the original photograph gives the video model a visual starting point. Google's Veo documentation specifically describes using reference images to guide scenes, characters, and objects, while also providing tools designed to improve consistency. �
Google DeepMind
For this reason, upload the clearest version of the original image available to you before entering the prompt.
The prompt then tells the model that the image should remain visually consistent and that only natural movement should be added.
Keeping the Cat Unchanged
One of the most important instructions in this project is:
“Keep the exact same orange tabby cat, same face, fur, body…”
This tells the generator that the cat is the main reference subject.
AI-generated videos can sometimes introduce unwanted changes when an animal moves. The fur pattern might change, the face may look different, or the body proportions may become inconsistent.
Using a clear reference image and explicitly requesting consistency can help reduce those unwanted changes. Google's image prompting guidance also recommends using clear reference images when maintaining subject consistency is important. �
Google DeepMind
Keeping the Mouse and Basket Consistent
The same principle applies to the mouse and basket.
The prompt says:
“Keep every character and object exactly the same throughout the entire video.”
This is useful because the mouse is an important part of the story. The basket is also important because it creates the visual idea of a possible trap.
The cheese should remain in its original location rather than suddenly moving around the floor.
The objective is not to create a completely new scene. The objective is to animate the scene that already exists.
Creating Natural Cat Movement
The cat should not suddenly jump, run, or make exaggerated movements.
Instead, the prompt requests:
Subtle breathing
Natural blinking
Gentle ear movement
Small whisker movement
Gentle paw movement
Natural body movement
These small details can make the final video feel more believable.
The cat's eyes can remain focused on the mouse while its ears react subtly to the mouse's movement. This creates a natural sense of attention without requiring complicated action.
Google's Veo prompt guide recommends describing the subject, action, setting, lighting, camera movement, and visual style when greater control over the generated video is desired. �
Google DeepMind
The Mouse's Movement
The mouse should also move slowly.
A useful sequence is for the mouse to cautiously approach the cheese and then pause after noticing the cat.
This creates a simple story:
The mouse sees the cheese → moves toward it → notices the cat → pauses.
The cat then watches the mouse.
This is enough for a 10-second clip without requiring complicated choreography.
Cinematic Camera Movement
The prompt uses a:
“very slow cinematic camera push-in.”
This means the camera gradually moves closer to the action.
A slow push-in can create suspense while keeping the original composition recognizable.
Avoid asking the camera to pan, rotate, zoom rapidly, and move backward at the same time. Too many camera instructions can make a short video unnecessarily complicated.
Google's current Veo documentation includes camera controls that allow creators to specify framing and camera movement. �
Google DeepMind
Lighting and Realism
The original image has a bright indoor kitchen with warm lighting and a wooden floor.
The prompt therefore asks for:
“Ultra-realistic fur, natural physics, realistic shadows and reflections.”
These details are important because the scene contains several surfaces and objects that should respond naturally to movement and light.
The wooden floor should retain its reflections. The cat's fur should respond subtly to movement. The basket should remain solid and stationary.
Realistic physics and visual consistency are among the capabilities Google highlights for its current Veo video-generation systems. �
Google DeepMind
Why the Video Should Be 10 Seconds
Ten seconds is a useful length for this concept because the entire story can happen quickly.
A possible timeline would be:
0–2 seconds:
The cat watches the mouse while remaining beside the basket.
2–4 seconds:
The mouse begins moving slowly toward the cheese.
4–6 seconds:
The cat subtly moves its paw and follows the mouse with its eyes.
6–8 seconds:
The mouse notices the cat and pauses.
8–10 seconds:
The camera slowly moves closer, ending on the cat, mouse, basket, and cheese.
This creates a beginning, development, and ending while keeping the movement controlled.
Vertical 9:16 Format
The prompt specifies:
“vertical 9:16.”
This format is useful when the final video is intended for mobile-focused platforms and short-form video content.
Keeping the original composition centered is particularly important because the cat, basket, mouse, and cheese all need to remain visible.
If your AI video tool provides an aspect-ratio option separately, select the desired format there as well.
Optional Sound Design
If the video generator supports audio, you can add subtle environmental sounds rather than loud effects.
For example:
Audio: quiet kitchen ambience, soft paw movements on the wooden floor, subtle mouse footsteps, gentle room ambience, and a light playful cinematic background sound.
Current Veo systems support native audio generation, including ambient sounds and sound effects. �
Google DeepMind
Keeping the audio subtle can help preserve the calm and playful feeling of the scene.
What to Do If the AI Changes the Image
If the first generation changes the cat's face, fur, basket, mouse, or cheese, strengthen the consistency instructions.
You can add:
Preserve the exact visual identity and position of every original subject and object. Do not redesign, replace, remove, or add anything. Animate only the natural movement of the existing subjects.
If the camera moves too much, use:
Keep the camera almost stationary with an extremely slow push-in.
If the cat moves too quickly, use:
The cat moves very slowly and naturally with minimal body movement.
If the mouse becomes too active, use:
The mouse makes only small, cautious movements toward the cheese.
These adjustments allow you to refine the result without completely rewriting the original prompt.
Creating the Video With Google Flow
Google Flow is one current option for AI-assisted video creation. Google's official Flow page says it supports tools for creating videos and lists capabilities including Frames to Video, Ingredients to Video, Video Extension, and video-to-video editing. �
Google Labs
You can learn more here:
Google Flow�
Google also provides an official Veo prompting guide with advice on camera movement, lighting, characters, locations, actions, and visual style. �
Google DeepMind
Google Veo Prompt Guide�
Final Thoughts
Turning a still photograph into a realistic 10-second video works best when the prompt clearly explains what should move and what should remain unchanged.
For this cat-and-mouse scene, the most important instruction is to preserve the original image. The orange tabby cat, mouse, woven basket, cheese, kitchen, colors, lighting, and composition should remain consistent while only subtle movements are introduced.
The cat can watch the mouse, the mouse can cautiously approach the cheese, and the camera can slowly push closer to create a playful moment of suspense.
A carefully written prompt gives the AI video generator a much clearer understanding of the desired result. Google's own prompting guidance recommends specifying elements such as the subject, action, setting, lighting, camera movement, and style when you want greater control over the generated video. �
0 Comments