1 / 6
Muse Cam: Raspberry Pi Digital Camera — image 1
1 / 6

Muse Cam: Raspberry Pi Digital Camera

Order a print of this model

We calculate the final price from your selected file. You review and accept it before payment.

Saved on this device.

Handheld Raspberry Pi camera that restyles every photo with AI. Six printed parts, magnetic doors, touchscreen, battery.

Rating
0 / 5
Likes
0
Downloads
1
Makes
0
Views
21
Reviews
0

Updated at source:

Model information

License
Creative Commons — Attribution

Muse Cam is a camera I designed around a Raspberry Pi 3B+. Pick a style on the touchscreen, such as Tiny Clay World, Found in 1997, Anime Cel, or Snow Globe, and press the shutter. The camera sends the photo to an image model and shows the restyled version on its screen. Every original and every result stays in the on-camera gallery.

It runs on its own: no server and no account with me, just your own API key for Meta's Muse Image model or OpenAI's image model. Each photo is one image edit billed to your API account.

Features

  • 30 built-in styles, picked from a rail on the touchscreen

  • 4.3" 800×480 capacitive touchscreen

  • Physical shutter button

  • PiSugar 3 Plus battery (5000 mAh), with the charge level on screen

  • Speaker for sound cues

  • Magnetic sliding doors: the side door slides forward to expose the Pi's USB and Ethernet ports

Printed parts (print one of each)

Part

Size on the bed (mm)

What it does

body_main

133 × 83 × 67.7 tall

Main housing: display, Pi and battery, amplifier, shutter. Prints display face down with the open front up

body_front

129 × 79 × 10.3 tall

Front cover; holds the lens assembly and speaker

door_side

79 × 60 × 5.3 tall

Side access door

door_top

64 × 48 × 5.3 tall

Top access door

lens_back

55 × 55 × 6.9 tall

Camera carrier and mounting flange

lens_front

45 × 45 × 6.0 tall

Front half of the camera sandwich

Every part fits a 180 × 180 mm bed. The STLs are already oriented for printing. The two-tone look in the photos comes from printing parts in different colors, so you don't need a multi-material printer.

Print settings

  • Material: PLA

  • Nozzle / layer height: 0.2

  • Walls / infill: 20%

  • Supports: body_main only, for back of where screen is inserted

  • Print time / filament: ~3-8 hours for all parts depending on if printed at once or individually

Clean up the magnet recesses before assembly. If one hole is tight, fix that hole; don't scale the whole model.

What else you need

  • Raspberry Pi 3 Model B+

  • Raspberry Pi Camera Module 3 (standard, not Wide or NoIR)

  • Waveshare 4.3" DSI touchscreen, 800×480

  • PiSugar 3 Plus with battery

  • MAX98357A I²S amplifier and a 3 W speaker

  • 10 mm momentary push button

  • 4 round magnets, 8 mm × 1.5 mm

  • M3 screws: 4 × M3×4, 9 × M3×6, 2 × M3×8 (the M2.5 screws come with the display and the PiSugar)

  • microSD card, medium CA glue, double-sided tape

The full parts list with links is on GitHub.

Assembly overview

About 2 hours once the parts are printed.

  1. Glue in the four magnets: two in the body and one in each door. Check that the pairs attract before gluing.

  2. Screw two M3×4 stop screws into the inside of each door.

  3. Mount the display, amplifier, and shutter in body_main, and wire the amplifier and shutter to the GPIO header.

  4. Connect both ribbon cables, then screw the Pi and PiSugar stack into the body.

  5. Sandwich the camera between lens_back and lens_front, then mount the lens assembly and speaker to body_front.

  6. Install the software and test everything before closing the camera.

  7. Close the front cover with four M3×6 screws.

The illustrated build guide has a photo for every step, and the wiring reference has every pin.

Software

Flash Raspberry Pi OS Lite (Legacy, 64-bit) with Wi-Fi and SSH, then run one command on the Pi and paste your API key when it asks:

curl -fsSL https://raw.githubusercontent.com/danny-hines/muse-cam/main/scripts/install-device.sh | sudo bash -s -- --profile pi3bplus-cam3-dsi43 --standalone

Add --provider openai to use an OpenAI key instead. Source code: github.com/danny-hines/muse-cam (MIT licensed).

License

The enclosure, build guide, and photos are CC BY 4.0. The software is MIT.

Tags