BioHue

BioHue is an application that analyzes images based on their color properties and classifies them into predefined categories. It consists of a Python-FastAPI backend, a React-Next.js frontend, and a MongoDB database for storing user data and analysis history.

Key Features

• User Authentication: Users can register and log in to upload images and view their analysis history.

• Image Analysis Workflow:

  • Users select a substrate and upload an image.
  • The backend processes the image and classifies it as Positive, Negative, or Moderate based on predefined thresholds.

• History Tracking: Analysis results are stored in MongoDB, allowing users to view past results.

Algorithm Breakdown

OpenCV is used for image analysis, following these steps:

1. Extracting the Prominent Region

   • Image Decoding: The uploaded image is converted from bytes into an OpenCV matrix.
   • Saturation-Based Region Detection:
     • The image is converted to HSV color space, and a binary mask highlights areas with high saturation.
     • Morphological operations (opening and closing) refine the mask by removing noise and filling gaps.
   • Contour Selection:
     • The algorithm detects contours in the binary mask and selects the largest one.
     • If the detected region is too small compared to the total image area, it is ignored.
   • Glare Removal:
     • Pixels with brightness exceeding a glare threshold are identified.
     • The average color of surrounding non-glare pixels is computed and used to replace glare-affected areas.
   • Circular Cropping:
     • A bounding box is drawn around the selected region, and a circular mask is applied.
     • This ensures consistency in shape and removes unnecessary background regions.

2. Computing the Metric

   • The extracted region’s RGB channels are separated.
   • A predefined mathematical formula (specific to the selected substrate) is applied to compute a numeric metric which is then classified.