Skip to content

Latest commit

 

History

History
18 lines (12 loc) · 905 Bytes

README.md

File metadata and controls

18 lines (12 loc) · 905 Bytes

Detection-of-Indian-Cars-with-brand-and-model

Developed a detection system which could recognize the brand and model of various vehicles in India

Team Members

  1. Praphul Singh - Undergrad, IIT Kanpur, Department of Electrical Engineering
  2. Preetham Paul Sunkari - Undergrad, IIT Kanpur, Department of Chemical Engineering
  3. Harshit Saini - Undergrad, IIT Kanpur, Department of Chemical Engineering
  4. Aadrish Sharma - Undergrad, IIT Kanpur, Department of Material Science & Engineering

Works

  1. We collected and filtered our own dataset consisting of images of cars and logos of 14 brands in India.
  2. Retrained the inception-v3 model separately for the brands and stored their weights giving an accuracy of approximately 80% for each brand.
  3. Used Hog filter and SVM to train the logos achieving an accuracy of 88%.
  4. Used sliding windows for recognizing logos of cars in a given image.