st.camera_input

To read the image file buffer as bytes, you can use getvalue() on the UploadedFile object.

import streamlit as st

img_file_buffer = st.camera_input("Take a picture")

if img_file_buffer is not None:
    # To read image file buffer as bytes:
    bytes_data = img_file_buffer.getvalue()
    # Check the type of bytes_data:
    # Should output: <class 'bytes'>
    st.write(type(bytes_data))

`st.camera_input` returns an object of the `UploadedFile` class, which a subclass of BytesIO. Therefore it is a "file-like" object. This means you can pass it anywhere where a file is expected, similar to `st.file_uploader`.

Image processing examples

You can use the output of st.camera_input for various downstream tasks, including image processing. Below, we demonstrate how to use the st.camera_input widget with popular image and data processing libraries such as Pillow, NumPy, OpenCV, TensorFlow, torchvision, and PyTorch.

While we provide examples for the most popular use-cases and libraries, you are welcome to adapt these examples to your own needs and favorite libraries.

Pillow (PIL) and NumPy

Ensure you have installed Pillow and NumPy.

To read the image file buffer as a PIL Image and convert it to a NumPy array:

import streamlit as st
from PIL import Image
import numpy as np

img_file_buffer = st.camera_input("Take a picture")

if img_file_buffer is not None:
    # To read image file buffer as a PIL Image:
    img = Image.open(img_file_buffer)

    # To convert PIL Image to numpy array:
    img_array = np.array(img)

    # Check the type of img_array:
    # Should output: <class 'numpy.ndarray'>
    st.write(type(img_array))

    # Check the shape of img_array:
    # Should output shape: (height, width, channels)
    st.write(img_array.shape)

OpenCV (cv2)

Ensure you have installed OpenCV and NumPy.

To read the image file buffer with OpenCV:

import streamlit as st
import cv2
import numpy as np

img_file_buffer = st.camera_input("Take a picture")

if img_file_buffer is not None:
    # To read image file buffer with OpenCV:
    bytes_data = img_file_buffer.getvalue()
    cv2_img = cv2.imdecode(np.frombuffer(bytes_data, np.uint8), cv2.IMREAD_COLOR)

    # Check the type of cv2_img:
    # Should output: <class 'numpy.ndarray'>
    st.write(type(cv2_img))

    # Check the shape of cv2_img:
    # Should output shape: (height, width, channels)
    st.write(cv2_img.shape)

TensorFlow

Ensure you have installed TensorFlow.

To read the image file buffer as a 3 dimensional uint8 tensor with TensorFlow:

import streamlit as st
import tensorflow as tf

img_file_buffer = st.camera_input("Take a picture")

if img_file_buffer is not None:
    # To read image file buffer as a 3D uint8 tensor with TensorFlow:
    bytes_data = img_file_buffer.getvalue()
    img_tensor = tf.io.decode_image(bytes_data, channels=3)

    # Check the type of img_tensor:
    # Should output: <class 'tensorflow.python.framework.ops.EagerTensor'>
    st.write(type(img_tensor))

    # Check the shape of img_tensor:
    # Should output shape: (height, width, channels)
    st.write(img_tensor.shape)

Torchvision

Ensure you have installed Torchvision (it is not bundled with PyTorch) and PyTorch.

To read the image file buffer as a 3 dimensional uint8 tensor with torchvision.io:

import streamlit as st
import torch
import torchvision

img_file_buffer = st.camera_input("Take a picture")

if img_file_buffer is not None:
    # To read image file buffer as a 3D uint8 tensor with `torchvision.io`:
    bytes_data = img_file_buffer.getvalue()
    torch_img = torchvision.io.decode_image(
        torch.frombuffer(bytes_data, dtype=torch.uint8)
    )

    # Check the type of torch_img:
    # Should output: <class 'torch.Tensor'>
    st.write(type(torch_img))

    # Check the shape of torch_img:
    # Should output shape: torch.Size([channels, height, width])
    st.write(torch_img.shape)

PyTorch

Ensure you have installed PyTorch and NumPy.

To read the image file buffer as a 3 dimensional uint8 tensor with PyTorch:

import streamlit as st
import torch
import numpy as np

img_file_buffer = st.camera_input("Take a picture")

if img_file_buffer is not None:
    # To read image file buffer as a 3D uint8 tensor with PyTorch:
    bytes_data = img_file_buffer.getvalue()
    torch_img = torch.ops.image.decode_image(
        torch.from_numpy(np.frombuffer(bytes_data, np.uint8)), 3
    )

    # Check the type of torch_img:
    # Should output: <class 'torch.Tensor'>
    st.write(type(torch_img))

    # Check the shape of torch_img:
    # Should output shape: torch.Size([channels, height, width])
    st.write(torch_img.shape)