Skip to main contentSkip to navigation
MachineHack Gen AI Logo
Cypher 2024 | Reveal the Concealed Logos Logo

Cypher2024|Reveal
theConcealedLogos

Expired
Start: September 23, 2024Ends: October 7, 2024
Participants
88
Time Left
Ended
Subs/day
12
Challenge Overview

Welcome to Cypher 2024 - India's Biggest AI Conference! 
To buck up your experience we are organizing a new interesting hackathon around working with image datasets.
 

This Hackathon is open to our community now!
 

Problem Statement:

The objective is to identify those images from the Image folder within the dataset that contain the machinehack logo hidden in some form either rotated slightly or a bit transparent within the main image.  

 

Dataset Description:

  • Images: Folder with all the image files.
  • Logo: Folder containing the MachineHack logos, tagged within the image files.
  • Train.csv: The file with image names and corresponding labels from the Image folder that can be utilized for training the model pipeline.
  • Test.csv: The image files from the Image folder for which you will generate outcomes.
  • Submission.csv: The format for submitting your predictions. The row sequence of outputs should match the sequence in the Test.csv file.

Participation and Benefits:

  • Skill Level: This challenge is perfect for participants with experience in image processing, computer vision, and deep learning techniques.
  • Community Engagement: Join our dedicated Telegram group to connect with fellow participants, share insights, and seek advice.
  • Recognition: All participants will receive a MachineHack certificate, and top performers will be showcased on the leaderboard.

Submission and Evaluation:

  • Submission Format: Submit your predictions in the provided submission.csv file.
  • Evaluation Metric: Submissions will be evaluated based on the Accuracy metric.
  • Leaderboard: Track your performance on the leaderboard and aim for the top positions.

How to Approach the Challenge:

  • Data Preprocessing: Clean the images by reducing blur, correcting rotations, and applying filters or transformations to improve clarity for logo detection.
  • Feature Engineering: Use techniques such as edge detection, feature extraction (SIFT, SURF), and convolutional neural networks (CNN) to identify logos within blurred and rotated images.
  • Modeling Techniques: Experiment with traditional image processing techniques like HOG (Histogram of Oriented Gradients), advanced deep learning models like CNNs or transfer learning models such as ResNet, VGG, or EfficientNet.
  • Validation and Tuning: Apply data augmentation and cross-validation to enhance model performance, and fine-tune hyperparameters to optimize results.

Getting Started:

  • Register Now: Ensure you're registered to participate and receive all necessary updates.
  • Download the Dataset: Access the dataset from the MachineHack platform to begin working on your solution.
  • Join the Community: Connect with fellow participants and mentors via our Telegram group for support and collaboration.

Support and Resources:

For any questions or assistance, contact our support team at support@machinehack.com. Stay updated by subscribing to our newsletter for the latest news and announcements.

We’re excited to see your innovative approaches to identifying the logos within the image files. Good luck and happy hacking! 🚀

Problem Statement

This challenge focuses on building advanced machine learning models to solve real-world problems. Participants will work with carefully curated datasets and compete to achieve the best performance metrics.

Target Column: labelled
Metric: accuracy_score
Level: Intermediate
Submissions: 12/day
Top Submissions

No leaderboard data available

Check back later for updates

Cypher 2024 | Reveal the Concealed Logos

Registration is open

Similar Challenges

Discover similar AI and data science competitions

No sponsored hackathons available at the moment.

Never Miss a Hackathon

Get notified about new AI hackathons, data science competitions, and exclusive opportunities. Join 50,000+ developers staying ahead of the curve.

No spam, unsubscribe at any time. We respect your privacy.