1: Upload an Image and ask anything- Literally anything!

You can upload an image by attaching image file directly to ChatGPT. If you are using mobile app, you can also take a picture and attach it to ChatGPT directly.

Example: Create a book catalogue from a library shelf picture

In this example, we will attach an image of a book shelf and ask ChatGPT to create a catalogue of books in the shelf and analyze them further.

Prompt : You are a librarian, analyze the books in the book shelf and provide a catalogue of books along with a brief description.

As you can see, ChatGPT was successfully able to create a catalogue of all the books ( with a brief description) just from a single picture.

Example: Analyze a random picture and get more details

In the below example we have uploaded a picture ( Which is a book cover ) and asked ChatGPT to provide more details. Here ChatGPT was able to browse the internet and provide more details on the book.

Example: Translate the text in the picture into a different language

In the below example we take a picture of a sign board in Chinese and ask ChatGPT to translate. As you can see , ChatGPT not only extracted the text ( via OCR), it identified the language ( Chinese) and translated it into English.

This is quite helpful, especially when you are travelling and you need to navigate the host country language ( Signboards, etc.)

2. Create a picture book with a specific character

In this example we will demonstrate the following
1. Create a specific character in ChatGPT
2. Create a unique id for the character (Generative ID)
3. Use the same character ( with the unique Generative ID) to create a series of picture stories (Which you can convert into a story book and publish)

Prompt 1: You are a children book illustrator. Create a cute 6 year old girl with red hair and glowing skin with cheerful facial expression. The girl is wearing a colorful dress with a unicorn design.

ChatGPT successfully creates an image of a girl as per instruction.

Prompt 2: Give me the Generative ID of the girl character

This step is very important. Generative ID is the unique ID that ChatGPT assigns to a image it creates. This is like your passport number or national identity number which uniquely identifies you as a person. As you see ChatGPT has provided a unique ID for the image ( Generative ID LoQ9RZzFROH236ur). We will use this unique ID henceforth to create various images of the same character in ChatGPT.

Prompt 3: Use the girl character you have created with Generative ID LoQ9RZzFROH236ur. Draw a image with her doing yoga.

As you can see ChatGPT has taken the same character and created a illustration with her doing yoga. If you don’t specify the Generative ID, ChatGPT might change the character whie creating the follow up illustrations.

Prompt 4: Use the girl character you have created with Generative ID LoQ9RZzFROH236ur. Draw a image with her dancing.

As you can see ChatGPT has taken the same character and created a illustration with her dancing.

You can continue further creating illustrations with this technique covering multiple characters. You can then ask ChatGPT to create a PPT or downloadable link of the images, which you can use to create your own picture book

3. Upload a video and ask ChatGPT to analyze the content.

Example: Create a catalogue of toys form a video of toy shelf

In this example we will upload a short video of a kid’s toy shelf. We will then ask ChatGPT to analyze the toy shelf and provide a catalogue of toys the kid plays with.

Prompt : Analyze the video of a kid toy shelf (attached) and create a catalogue of toys you see.
As you can see ChatGPT successfully analyzes the video and creates a catalogue of toys. The way it does this is it it takes frames (snapshot) of the video at an interval of 60 seconds ( Frame 0, Frame 60, Frame 120..) and analyzes each frame.

Example: Extract various characters from a video

In this example we will upload a short video of a flower garden and ask ChatGPT to extract all unique flowers, from the video, it sees.

Prompt : Analyze the video ( Attached) of the garden. Extract each unique flower frame as a separate image.

As you can see ChatGPT has identified set of unique flowers from the video and have extracted images of them ( frame from the video).