Skip to content

ObjectCount.md

Latest commit

 

History

History
131 lines (91 loc) · 3.65 KB

ObjectCount.md

File metadata and controls

131 lines (91 loc) · 3.65 KB

Internship Assignment: Image-Based Object Counting using Python

Objective

Assess the candidate’s ability to work with Python, image processing, and basic computer vision concepts by building an application that analyzes an image and returns the count of distinct objects arranged in stacks or blocks.

Assignment Overview

You are given an image containing multiple stacked objects (for example, stacks of clothes arranged in rows and columns such as 3×3).
The task is to analyze the image, ignore noise, identify similar objects based on shape, size, or block structure, and return the total count of distinct objects.

The application should also provide a simple UI where a user can upload an image and trigger the object-counting process.

Problem Statement

  • Input image may contain:
    • Objects stacked vertically or horizontally
    • Similar-looking objects (colors may repeat)
    • Noise around or inside the image
  • Example:

    • If the image contains 9 stacks of cloths arranged in a 3×3 grid, the output should be:

      Total distinct object count: 9

The logic should rely on image analysis, not manual hardcoding.

Functional Requirements

1. Image Upload

  • Provide a UI input to upload an image.
  • Display the uploaded image thumbnail on the screen.
  • Validate that the uploaded file is a valid image format.

2. Object Detection & Counting

  • Analyze the image to:
    • Detect object boundaries or blocks
    • Identify similarity using width, height, contours, or block-based segmentation
  • Ignore:
    • Small noisy regions
    • Irrelevant artifacts
  • Return the total number of distinct objects detected.

3. Count Trigger

  • A “Count Objects” button that:
    • Processes the uploaded image
    • Displays the final object count clearly

Technical Constraints

  • Must be implemented using Python
  • Allowed libraries:
    • OpenCV
    • NumPy
    • Any open-source image processing or ML library
    • Open-source LLMs (optional, if used for assistance or reasoning)
  • The solution should not rely on manual marking of objects.

UI Requirements

  • Simple UI (any one is acceptable):
    • Streamlit
    • Flask + HTML
    • Gradio
  • UI must include:
    • Image upload input
    • Uploaded image preview
    • Button to start object counting
    • Clear display of the result

Suggested Approach (Not Mandatory)

  • Image preprocessing (grayscale, blur, thresholding)
  • Noise removal using morphological operations
  • Contour detection or block segmentation
  • Filtering contours based on size/aspect ratio
  • Counting valid detected objects

Submission Instructions

  • Complete the assignment within one week
  • Upload the project to your own GitHub repository
  • Include a README.md with:
    • Problem explanation
    • Approach and algorithm used
    • Libraries used
    • How to run the project
    • Sample input and output screenshots

Bonus Points

  • Show bounding boxes around detected objects
  • Display intermediate steps (noise removal, segmentation)
  • Support both horizontal and vertical stack layouts
  • Ability to handle different image sizes gracefully

Evaluation Criteria

  • Correctness: Accurate object counting
  • Image Processing Logic: Proper noise handling and segmentation
  • Code Quality: Readable, modular, and well-commented Python code
  • UI Usability: Clear and simple user flow
  • Documentation: Clear README and explanation

Good luck! This assignment tests how well you can make machines “see” structure inside visual chaos 📷🧠

© 2026 RecursiveZero, All rights reserved.