Redraw and Correct a Figure from a Photo
A clean, correctly labelled redraw of a messy or faint textbook figure you photograph.
Makes images - needs an image AI · Best tool: ChatGPT or Gemini (image)
✅ How to use this effectively
- 1. Open an image-capable AI chat.
- 2. Paste the prompt below.
- 3. Attach a clear photo of the figure.
- 4. Send and check every label.
- 5. Download the redrawn image.
🔧 If it is not right, reply with this: If a value was misread from the photo, reply: "The side reads 7.5 cm, not 75. Correct it and redraw the figure cleanly."
THE PROMPT
ROLE: Act as a draughtsman who reads a photographed maths figure and redraws it cleanly and accurately. CONTEXT: I teach [CLASS/GRADE] under [BOARD or EXAM e.g. CBSE / ICSE / State Board]. My textbook or board figure is faint, crowded or skewed, and I want a clean version with the same information. YOUR TASK: 1. Read the attached photo carefully and identify the figure: shape, points, lines, angles, and every label and measurement. 2. List back to me everything you can read from the photo, so I can correct any misread value before you draw. 3. Redraw the same figure neatly: straight lines, proper proportions, plain white background. 4. Keep every label and given value from the original, placed clearly. 5. Add standard marks the original implies: tick marks for equal sides, arcs for equal angles, a small square for right angles. INPUT: I have attached a photo of the figure. If any text in the photo is unclear, ask me rather than guessing. HOW TO WORK IT OUT: Transcribe the figure's data first and present it as a short list (points, sides with lengths, angles, special marks). Do not invent anything the photo does not show; where the photo is unreadable, write "unclear - please confirm" and wait. Once values are confirmed, VERIFY two ways before finalising: check the redrawn figure includes every item on your transcribed list, and check that the stated measurements are internally consistent (flag any contradiction). Keep all labels legible when projected. OUTPUT FORMAT: 1. WHAT I READ: a short list of all points, sides, angles, values and marks from the photo, with any "unclear" items flagged. 2. THE REDRAWN FIGURE: one clean labelled image matching the confirmed data. 3. CHECK: one line confirming every transcribed item appears in the redraw. RULES: Never invent measurements not visible in the photo. Flag unclear items instead of guessing. Straight lines, correct marks, plain white background, legible labels. SIGNATURE: Sign the bottom-right of every page with an elegant handwritten signature "[YOUR NAME]" with a small underline flourish.