/streamlit-convolution

Primary LanguagePython

Streamlit Kernel Computation Demo

Original author: Jan Korytář

This Streamlit demo showcases kernel computation using sliders for customization and visualization.

Demo Image

Libraries

The following libraries are imported:

  • streamlit as st
  • pandas as pd
  • numpy as np
  • correlate2d from scipy.signal

Sidebar Customization

  • Kernel Customization: Users can adjust kernel dimensions using a slider in the sidebar.
    • If the kernel is not initialized or dimensions change, a random kernel with specified dimensions is created.
    • The kernel is displayed in a data editor for user interaction.

Main Interface

  • Title: "Simple kernel computation demo"
  • Kernel Settings Sidebar: Users can set array width, height, and indices using sliders.
  • Instructions: Users are prompted to select the upper-left corner of the kernel.

Visual Representation

  • The DataFrame displays array data with highlighted user-selected area.
  • DataFrame Display:
    • Columns and rows are highlighted based on user-selected indices.
  • Output Display:
    • Three columns are displayed:
      1. Chosen Area: User-selected array area
      2. Kernel Data: Kernel displayed in the data editor
      3. Combined Operation: Multiplication of chosen area and kernel, showing results

Convolution Result

  • Convolution: Utilizes correlate2d from scipy.signal to compute convolution result.
  • Convolution result is displayed in a DataFrame.