The Nano Banana trend and the Vintage Saree AI trend have stormed the internet, with thousands of users creating dreamy retro looks and hyper-realistic figurines of themselves. As more people jump on the bandwagon, here are a few tips and tricks to help you generate the best results.
Give specific prompts: The magic of these trends lies in how prompts are crafted for Google Gemini. Prompts should be detailed not only about clothing but also the background, angles, textures, and overall mood of the image. For example, instead of a vague prompt like “a black saree,” try “a black shimmering chiffon saree with a deep wall and dramatic shadow in the background.”
Provide a proper reference image: AI outputs are heavily influenced by the reference image you upload. Ensure the reference photo clearly shows your face, is well-lit, uncluttered, and not blurry. A good-quality reference makes the AI output look far more realistic and polished.
Also read: 5 tips to make your saree AI edit look like a real Bollywood poster
Keep fixing for a better result: If the first picture is not up to the mark, keep giving prompts to fix the issues and make a better image. With this, you can fix little details like the colour or background texture. Adding phrases like ‘soft shadow’ can lighten the shadow in the image.
{{/usCountry}}Keep fixing for a better result: If the first picture is not up to the mark, keep giving prompts to fix the issues and make a better image. With this, you can fix little details like the colour or background texture. Adding phrases like ‘soft shadow’ can lighten the shadow in the image.
{{/usCountry}}Experiment with actions: Users can also experiment with actions like making an image in a sitting position or with arms folded for a more dynamic and unique look. Further, you can experiment with objects like flowers in hair or making wavy hair.
Add elements: Google Gemini also allows you to add an element to an image, like a pair of sunglasses. Just give the prompt, ‘using the provided image of (subject of image), please (add/remove) (element) from the scene. You can further add how the change should be integrated.
Add keywords: Adding keywords in the prompts will guide the AI on how the image should turn out. There are several keywords out there, like dynamic, cinematic, retro, realistic, and vintage, depending on the overall mood of the image.
Also read: Gemini AI saree edits go viral on Instagram: Step-by-step guide to create retro look
Try different styles: Rather than sticking to the general realistic styles, users can experiment with different types of styles, like Pixar-style realism with glossy eyes and lifelike textures or making a fantasy sorcerer version of the Gemini Nano Banana holding a sword.
Don't overload prompts: Try not to overload the information in the prompt. Adding too many details can create confusion, and the image might not be generated as needed.
Add facial consistency prompts: While giving the prompt, remember to add instructions like ‘keep the facial features the same’ or ‘do not change the facial features and shape’. This will prevent Gemini from distorting the original facial features.
Specify facial expressions: AI images are also about showing emotions, which can only be integrated by giving specific instructions about expressions. Add instructions like ‘serene expressions’, ‘calm yet romantic’, intense and dramatic’, etc.